You’re a great Product Manager.

You know you’ll be an even better Product Leader. But...

Why is it so hard to make the leap from PM → Product Leader?

Hard Truth #1:
Opportunities for advancement in product management are scarce. Product teams are typically smaller, and when leadership roles open, the competition is fierce. You need to be primed and ready for when your chance comes.

Hard Truth #2:
The skills that make you a stellar PM won't necessarily make you a stellar Product Leader. Leading 100 people isn't the same as leading 10; it requires a significant evolution of your skill set.

So, what skills do you need? How do you acquire them?

That's where I come in!

I’ve been building products for over 25 years, and have grown from leading a small team of 5 to leading a 300+ org at a top-tier company. I became a Product Leader by navigating the very same obstacles facing you today.  

My mission is to help great PMs like yourself transform into outstanding Product Leaders.

You can do it! There are immediate steps you can take starting today to set yourself on this path.

Follow me here on LinkedIn to discover how.


- - - - -
🔥 If you *really* want to fast-track your career, visit: amplifypm.com/transform

Roshan Gupta

Product managers, Backstreet Boys, and 4-year-olds have one thing in common.


"Tell me why."

Except 4-year-olds eventually stop asking.
Backstreet Boys made millions asking.
PMs who stop asking get passed over.

The best PM I've worked with asked "why" so relentlessly that it was borderline annoying.

- Why is this on the roadmap?
- Why does this metric matter?
- Why are users dropping off here and not there?
- Why are we building this now?

Not to be difficult. But to make sure the team was solving the right problem before writing a single line of code.

Product leaders interrogate the feature request until it either becomes a better idea or dies.

"Tell me why" isn't a junior behavior. It's a leadership behavior that most people unlearn because it feels uncomfortable in rooms with strong opinions.
Don't unlearn it.

Backstreet Boys are still touring. The 4-year-old is onto something. And the best PMs never stopped asking.

If this made you think of someone on your team, share it.

Follow me to go from PM → Product Leader 🪜 | Founder of AmplifyPM (ex-Google, Meta, Yahoo, startups) | I grew from leading a team of 5 to leading a 300+ org. I'll show you how to do the same.

Most PMs spend their careers optimizing for things no one will remember.


→ PMs obsess over Jira tickets.
→ They count meetings.
→ They replay how many features shipped last quarter.
→ They carry guilt over tech debt they had to defer.

All of that feels important in the moment. None of it survives time.

What actually sticks is different.

→ People remember whether you could explain where the product was going when things were unclear.
→ They remember if you could connect scattered ideas into a direction that made sense.
→ They remember whether your solution felt inevitable once you explained it.
→ They remember how confident they felt walking out of the room.

That’s the quiet difference between being busy and being influential.

Early in my career, I thought execution would speak for itself.
Later, I learned that execution without narrative just becomes background noise.

The PMs who become product leaders aren’t the ones who do more.
They’re the ones who help others see more clearly.

→ That’s what creates trust.
→ That’s what gets repeated when you’re not in the room.
→ That’s what people actually remember.

P.S. If you feel like your work is solid but your impact isn’t sticking, this is usually the missing layer. I now spend most of my time helping PMs turn strong thinking into clear, persuasive narratives that senior leaders trust. 

Register for Persuading People: The PM’s Edge in the Age of AI, and go from a PM to product leader.

Save 33% when you register today: https://lnkd.in/d7STtuG8

A Sr. PM just added $100K to their total comp by making decision makers feel how losing them would cost the company millions.

They did not do it by negotiating harder or by flashing a competing offer. They did it by controlling how the room felt about the cost of saying no.

I know what some of you are thinking: that's just a smart negotiation tactic.
But I've watched this exact skill show up at every inflection point in a PM's career, and it has nothing to do with negotiation.

When a roadmap gets funded over three others with better metrics, 

it's usually because one PM made the tradeoff feel real while everyone else reported numbers.

When a mid-level PM gets promoted ahead of peers with stronger execution records, 

it's almost always because they made leadership feel the consequence of not betting on them.

When a product decision gets made in a room with six strong opinions, 

the PM who shapes the outcome isn't usually the one with the most data; it's the one who made the cost of inaction impossible to ignore.

This is what I mean when I say storytelling is not a soft skill. It's the mechanism that helps PMs convert good work into real influence.

And the gap between PMs who have it and PMs who don't shows up everywhere - in
↳ Whose roadmap gets prioritized
↳ Whose promotion lands
↳ Whose recommendation gets acted on without a second meeting.

The market has already figured this out, that the PM who controls how a room feels about a decision controls the decision itself.

Every week you spend reporting instead of shaping belief is a week someone else is building that skill. Persuading People: The PM's Edge in the Age of AI has helped 150+ PMs close that gap in how they frame decisions, lead rooms, and earn bigger bets.

Register now if you want to go from a PM to a Product Leader: https://lnkd.in/dB7xfrvr

undefined undefined

PST (Maven)

Master the persuasion system I used to secure 9-digit budgets & Fortune 50 partnerships: amplifypm.com/maven

One-Off - 2026-01-29

amplifypm

Everyone can see I went from PM at Meta to GPM at Google. What nobody sees is that I went from leading a team of 10 to leading 300 people.

That second jump was the harder one.

With 10 people, I knew everyone.
I knew what kept them up at night, what motivated them, what made them move. My message never had to travel far to land.

At 300, everything I said passed through layers I couldn't control.

By the time my thinking reached the people doing the work, it had been filtered and reinterpreted so many times that what landed had nothing to do with what I intended.

Teams pulled in the wrong directions. Decisions got made without context. And I kept wondering why nothing was sticking.

I was still communicating like I led 10 people.

When your team is small, relationships carry your message. People trust you because they know you.
At 300, most of the org doesn't know you. Your message has to stand on its own, clear enough and compelling enough to survive every layer it passes through.

That's when I understood storytelling wasn't a soft skill.
It was the only thing that scaled.

Knowing what your audience fears, what they want, and what would make them move, and building every message around that.

The jump from PM to GPM looks like a title change from the outside. From the inside, it was learning to lead people who had no reason to trust me yet.

When I made that jump, there was no playbook for it. I had to figure it out on my own.

That's exactly why I built Persuading People: The PM's Edge in the Age of AI, so you don't have to. 150+ PMs have used it to earn trust faster and lead rooms with confidence.

Register here if you're ready to go from a PM to product leader: https://lnkd.in/dB7xfrvr

There's a version of PM work where you show up with all the right answers and still lose the room. I lived that version for longer than I'd like to admit.

I had spent years getting better at the craft.
I mastered sharper prioritization, cleaner PRDs, and tighter metrics. And none of that was wrong.

But when I look back at the moment my career actually shifted, it wasn't a framework that changed things.
It was learning how to make senior people feel confident in a room.

Not confident in me. Confident in the decision.

I'd walk in with data, a clear recommendation, a well-structured slide, and somehow the energy would stay flat.

Decisions would get deferred. I'd leave wondering what I missed.

What I was missing was the stakes.

The shift happened when I stopped thinking about how to explain my thinking and started thinking about what I needed the room to believe by the time I finished speaking.

At Google and Meta, the PMs who moved fast weren't always the ones with the best product instincts.

They were the ones who could walk into a room with three VPs, a half-baked idea, and enough clarity about the problem that everyone left feeling like the direction was obvious.

That's not a communication skill. That's a leadership skill that shows up in communication.

If you're doing everything right and still feeling stuck, it might not be your execution. It might be that nobody in your organization can feel the importance of what you're building, and that's on how you're framing it, not on the work itself.

That gap between shipping well and shaping belief is exactly what I built Persuading People: The PM's Edge in the Age of AI to close. 

More than 150 PMs have used it to sharpen how they frame decisions, lead rooms, and earn bigger bets. If that's the gap you're sitting in right now, register here: https://lnkd.in/dB7xfrvr

A lot of people will defend this. I won't.


The Uber CEO just told the world: answer my weekend emails immediately or you're out. "Don't come here if you want to coast."


I'm not against hard work. I've worked hard my entire career.
But there's a difference between high expectations and treating people like they have no other options.

I've watched this play out with teams for 26 years. I can tell you exactly how it ends.

There's a concept I use with clients: the golden goose and the golden egg.

Short-term thinkers obsess over the next egg. The next milestone. The next earnings report.

What they forget is the goose. The goose is your team.

Burn it out chasing short-term results, and you quietly destroy the very thing generating those results.

Here's the other irony: AI was supposed to give us back time.

Instead, the efficiency gains aren't going back to the people doing the work.
They're being used to shrink headcount and squeeze harder. 
That's a leadership choice. Not a technology requirement.

Trust doesn't show up in a quarterly report. It shows up or disappears over the years.

And a team performing out of fear produces very differently from a team performing out of belief.

But here's what these leaders are miscalculating: they're assuming workers have no other options. They might be surprised.

There's an idea Tim Ferriss calls the "tolerable middle" - when people are uncomfortable but not miserable enough to act, they stay indefinitely.

Push them past that line, and it becomes the catalyst they needed to build something, start something, or walk away to something better.

The same AI giving companies leverage is quietly giving workers new options. To freelance, build, and leave.

You don't have to endorse coasting to reject this model. High standards and genuine respect for people are not opposites.

It takes real courage to lead that way, especially when competitors are racing in the other direction.

But that's what separates leaders who build something lasting from those who just squeeze something dry.

If this made you think differently about how you lead or how you're being led, that's exactly the muscle I help PMs build inside Persuading People: The PM's Edge in the Age of AI. 

Register now to go from PM to Product Leader: https://lnkd.in/dB7xfrvr

Product managers, the executives sitting in your next presentation don't care about your data. (So stop obsessing over it entirely.)

They're deciding whether to trust you.

I learned this across my time as a PM at Meta and Google.

The PMs who consistently won executive rooms weren't the ones with the most rigorous analysis.
They were the ones who made executives feel like every possible concern had already been anticipated.

Here's what that looks like in practice.

Before building a single slide, the best PMs do one thing most skip: they write down every question their audience might ask.

The easy ones.
The skeptical ones.
The ones an excited exec asks.
The ones a hostile exec asks.

Then they group those questions into themes.
Each theme reveals what the presentation actually needs to answer.
The arguments practically write themselves from there.

Then comes the move most PMs never think to make.
Everything that doesn't make the final story, extra evidence, alternative angles, leftover questions, goes into an appendix.

When an executive raises a concern mid-presentation, and you jump straight to a prepared answer, this changes everything.

They stop evaluating your idea. They start trusting your judgment.

That trust follows you into every future meeting.

Most PMs walk in hoping to impress executives with data depth.
The ones who move fastest walk in knowing exactly what the room is worried about, and having already answered it before anyone speaks.

That's not a presentation skill. That's a reputation being built in real time.

P.S. If this reframed how you think about executive presentations, this is exactly what I teach inside Persuading People: The PM's Edge in the Age of AI.

150+ PMs have used it to sharpen how they frame decisions, lead rooms, and earn bigger bets.

If you are ready to go from a PM to a product leader, register here: https://lnkd.in/dB7xfrvr

If I had to bet it all, I'd bet on the PM who can explain a problem clearly over the PM who can solve it perfectly.

I know that sounds wrong.

But I've watched it play out too many times at Google and Meta to ignore it.

The PM who solves problems perfectly but can't communicate their thinking clearly gets questioned in every meeting, pulled into edge cases, and asked to "come back with more data."

They're brilliant. But they're ineffective.

Because as a leader, I don't have time to dig into the details. 
I can't afford to hope they got it right. 
Hope is not a strategy. I need proof. I need evidence. I need to trust them – and trust is the deciding factor when I'm choosing which PMs to bet on.

The PM who explains the problem so clearly that everyone in the room immediately understands what's at stake? They get trusted. They get resources. They get the next big bet.

They're not always the sharpest technically.
But they're always the clearest communicator.

Senior leadership doesn't have time to evaluate your solution in the room. They don't have the context you have. They haven't lived inside the problem for six weeks like you have.

What they're actually evaluating in real time is your judgment.

And judgment shows up in how you frame the problem. Not how you solved it.

If your framing is sharp, they trust the solution.
If your framing is muddy, it doesn't matter how good the solution is.

So if I had to bet it all – and in every promotion decision I've been part of, I did have to bet – I'd bet on the PM who makes the room smarter just by how they explain things.

Every single time.

Whom would you bet on?
--
P.S. I created the live "Persuading People: The PM's Edge in the Age of AI" cohort to help PMs communicate decisions so clearly that leadership trusts your judgment, so you get seen and rewarded. 

More than 100 PMs have already used it to accelerate their growth. Register now if you want to go from a product manager to product leader: https://lnkd.in/d7STtuG8

"It's just been 3 weeks into my PM journey, and I'm already 'that guy'."


I just bumped into this thread and couldn't scroll past it.

A new PM spent a week on their first roadmap presentation. They had vision, personas, a polished deck, and thought they nailed it.

Then as people were dropping off the Zoom, a mic stayed hot, and someone said, "That completely could have been an email."

As PMs, we have all been there.

But here's the thing nobody told this PM, and nobody tells most PMs early enough: A presentation isn't there to inform. It's there to transform.

If all you're doing is sharing facts and figures, send the email.
A presentation earns its place on the calendar only when something needs to change in the room.

Before your next presentation, answer these four questions:
→ What does my audience currently think or feel about this topic?
→ What do I need them to think, feel, or do differently when I'm done?
→ What is the single most important thing I need them to walk away with?
→ Is the gap between where they are and where I need them to be realistic in one session?

That last one matters more than most PMs realize.
Sometimes the goal isn't to get approval, it's just to open a mind that wasn't open before. That's still a transformation.

I'm sure the PM in that Reddit thread built a great deck. But they just never asked what they needed to change in the room.

That's the real job.

P.S. If this hit something, defining your audience's transformation before opening a deck is one of the things I teach inside Persuading People: The PM's Edge in the Age of AI.

150+ PMs have used it to sharpen how they frame decisions, lead rooms, and earn bigger bets. 

If you are ready to move your career, register here: https://lnkd.in/dB7xfrvr

You chased the PM title for years. But no one warned you what it would feel like once you got it.


You thought the hard part was breaking in.
The certifications, case studies, mock interviews, and side projects.

Then you land the role.
And suddenly you’re in meetings where the stakes are higher than anything you practiced for.

Engineers are waiting for clarity.
Stakeholders are demanding a roadmap you’re still forming in your head.
Leadership expects conviction.

And you’re thinking: Am I actually ready for this?

That feeling has a name. Imposter syndrome.
You don’t feel that way because you’re unqualified.

But because the role now demands judgment.

Here’s what I’ve seen work, both in my own career and mentoring PMs over the years.

→ Claim it

Stop pretending you’re fully confident. Growth feels unstable at first. Accept that you’re stretched.

→ Name it

Get specific. Are you unsure about the metric? The technical depth? The stakeholder dynamic? Vague anxiety feels overwhelming. Specific doubt becomes solvable.

→ Frame it

Turn fear into facts.
What decision actually needs to be made?
What data do you have?
What assumptions are you making?
What tradeoff are you avoiding?

→ Tame it

Most imposter syndrome is unstructured thinking under pressure. When you structure the decision, the “fraud” feeling shrinks.

The PMs who grow fastest aren’t the ones who feel zero doubt.
They’re the ones who learn how to think clearly while feeling it.

Confidence doesn’t come from titles.

It comes from being able to prepare your arguments, anticipate objections, and guide a room toward a decision even when you don’t have perfect certainty.

If you’re feeling that discomfort right now, it probably means you’re operating at a higher level than you were six months ago.

And that’s not fraud. That’s growth.

If imposter syndrome is showing up in senior rooms, it’s time to strengthen how you prepare, frame, and defend your decisions.

Join the next Persuading People cohort and build that muscle deliberately like 100+ PMs did.

If my course doesn’t change you, you get your tuition back. 

Last chance to save 33% when you register today https://lnkd.in/d7STtuG8

For years, I used to believe great PMs were better thinkers. 
But I was wrong.

They're better narrators.

After almost 2 decades as a PM and Group PM, mentoring 150+ PMs, I can tell within 10 minutes who is operating at the next level.

Not by looking at their roadmap or metrics. But from how they tell the story.

Average PM: We shipped X. It improved Y by 12%.

Strong PM: Three quarters ago, we were losing users at onboarding because of friction in step 2. We believed speed was the issue. It wasn't. It was trust. Here's what we changed. Here's what it unlocked.

Same outcome, but a completely different signal.

The PMs who move up consistently know how to:

→ Build a Story Map before they open a deck
→ Define the one transformation they want from their audience
→ Tap into what their audience actually cares about - not what they assume
→ Find the "Super Why" that motivates action, not just understanding
→ Open with a knowledge gap that makes the room lean in
→ Use emotional tension to make the status quo feel uncomfortable
→ Choose the right narrative structure for the right audience
→ Turn their appendix into a trust signal with executives
→ Deliver with the kind of conviction that makes belief contagious
→ Debrief after every pitch so their storytelling compounds over time

They don't just ship. They shape meaning.

And that's what executives are really evaluating.

This is why two PMs with identical execution histories end up on different trajectories.

One reports, the other reframes.
One explains what happened, the other defines why it mattered.

If you look at those 12 storytelling patterns carefully, they're not communication skills. They're leadership behaviors that look like storytelling.

That's the shift most PMs miss.

You don't become a product leader when someone gives you the title. You become one when you control the narrative before anyone asks you to.

If this resonates, save it.

And if you see yourself in 5 or more of those 12 signals, you're already playing a different game.

If you're reading this and realizing you've been shipping well but not shaping belief, that's exactly what I teach in Persuading People: The PM's Edge in the Age of AI.

More than 150 PMs have used it to sharpen how they frame decisions, lead rooms, and earn bigger bets.

If you're ready to move lanes, register here: https://lnkd.in/dB7xfrvr

I’m a PM on paper. But in meetings, I feel like an assistant.


I saw that line on Reddit this week.

It was from a PM who was exhausted.
The person wasn’t bad at their job.
They were shipping, their metrics were green, and roadmaps were full.

But in exec reviews, they’d present, then someone else would restate the problem in simpler terms, and that version would land.

That part is painful.

Because it makes you question yourself.

You think, “Didn’t I just say that?”

You did.

But you said it like a PM updating status.

They said it like someone setting stakes.

That Reddit thread also said, “When I stopped explaining what we built and started explaining what breaks if we don’t act, the room changed.”

Nothing about the work changed.
The framing did.

A lot of good PMs get stuck here.

They think influence comes from being thorough.
So they add more detail.
More context and slides.

But senior rooms don’t reward completeness.

They reward clarity.

And this is where AI makes it worse.

AI will help you generate perfect updates.
It will make your slides cleaner.
It will summarize your thinking beautifully.

Which means “clear documentation” is no longer a differentiator.

What’s left?

Judgment, framing, and timing.

The ability to walk into a room and say the thing that actually matters.

That’s not a communication trick. That’s a skill.

And most PMs only realize they’re missing it when they start feeling invisible.

If you’ve ever walked out of a meeting feeling invisible despite doing solid work, this is a skill worth building deliberately.

That’s exactly why I built Persuading People: The PM's Edge in the Age of AI
If you want to move from “update person” to “trusted voice,” register here: https://lnkd.in/dB7xfrvr

Senior PM and Group PM comp grew +25% and +13% in the last 2 years.
If you’re still “doing your job well and waiting for a promotion", you’re in the wrong lane.

That kind of jump doesn’t happen because everyone suddenly got better at writing PRDs.

It happens because companies are paying a premium for a different kind of PM.

Over the last decade, I’ve watched this shift up close.

There was a time when being reliable was enough.
If you shipped on time, kept stakeholders informed, and hit your metric.
That used to earn you the next title.

In 2026, promotion conversations sound different.

The questions aren’t: Did they deliver?

They’re:
↳ Would I trust them with a bigger bet?
↳ Can they handle ambiguity without escalating?
↳ Do they think in business tradeoffs, not feature requests?

The PMs moving up fastest are already operating one level above.

↳ They frame problems in terms of company impact, not backlog priority.
↳ They make tradeoffs explicit instead of smoothing them over.
↳ They drive alignment before the meeting, not inside it.
↳ They communicate with clarity that reduces anxiety in the room.

They don’t wait for the title to start acting like a product leader.
That’s the real shift.

If your plan is to “do great work and let it speak for itself,” you’re competing in a crowded tier.

At the top, clarity, judgment, and influence are the differentiators.

Comp growth is just the market’s way of pricing that in.

If you want to operate at the next level before anyone hands it to you, that’s exactly why I built Persuading People: The PM’s Edge in the Age of AI.

More than 150 PMs have used it to sharpen how they frame decisions, lead rooms, and earn bigger bets.

If you’re ready to move lanes, register here: https://lnkd.in/dB7xfrvr?

Netflix just posted a role paying $775,000. For someone who can write words and tell a story. If you're a PM wondering what leadership actually values, there's your answer.

Here's what's happening in the market right now.
Anthropic tripled its communications team last year.
OpenAI has open comms roles paying over $400,000.
Job postings on LinkedIn mentioning "storyteller" have doubled.

This isn't a coincidence.

AI is flooding every company with content. Decks, updates, summaries, briefs - all generated in seconds. All sounding exactly the same.

And in that sea of sameness, the person who can communicate with clarity and human conviction has become the scarcest resource in the building.

I saw this long before the market caught up.
At Google and Meta, the PMs who got the biggest bets weren't always the most technically gifted.

They were the ones who could walk into a room and make leadership feel the weight of a problem, before a single slide went up.

That skill didn't show up in job descriptions back then.
Now companies are paying $775,000 for it.

As a PM you have been trained to optimize for output.
Ship faster, build more, and move the metric.

But leadership doesn't promote the PM who builds the most.
They promote the PM who makes them confident in what's being built, and why it matters.

That confidence doesn't come from your roadmap.
It comes from your story.

AI can generate content. It cannot generate credibility.
And in 2026, that distinction is worth more than any technical skill you could add to your resume.

The hottest job in tech is weaving together a compelling story.
Make sure you're the PM who knows how to do this.

P.S. If you want to understand why companies are suddenly paying top dollar for communicators, not just builders, I did a deep dive on this in a Maven session: Persuading People: The Skill AI Can’t Replace.

In 30 minutes, I walk through how to structure arguments, frame tradeoffs, and anticipate objections in high-stakes rooms.

It’s available free here: https://lnkd.in/gNQxUqRp

PMs when someone finally mentions product vision, because we've spent the last three months being asked why the button isn't blue yet.

Every PM knows this feeling.

You got into this role to solve real problems, shape strategy, and build something that actually matters.

Then Monday arrives, and your calendar is back-to-back with spec reviews, stakeholder complaints, and a very heated Slack thread about font sizes.

The dangerous part isn't the distraction. It's how easy it is to start believing that's the job.

The PMs I've seen stall out aren't lazy or unambitious.

They just slowly stopped asking the bigger questions: why are we building this, what does winning actually look like, what would we stop doing if we were serious about the strategy, because nobody around them was asking those questions either.

The best PMs I've worked with never fully let go of that.
Even when they were deep in execution, when the quarter was on fire, they kept one hand on the vision.

It's what made their decisions faster, their tradeoffs cleaner, and their roadmaps feel like a direction instead of a list.

The button color matters. But only if you know why the button should exist in the first place.

If you want to operate that way, where vision and execution aren't opposites but the same muscle, that's exactly what I focus on inside Persuading People: The PM's Edge in the Age of AI.

Register here if you want to go from a PM to a product leader: https://lnkd.in/dB7xfrvr

As a GPM at Google, I've sat in 100s of product reviews. And I can tell you the exact moment a PM loses the room.

It's not when the data is weak, or when the roadmap is half-baked.

It's when they stand up and spend 20 minutes talking about what they built, and zero minutes talking about what someone felt.

I've watched talented PMs walk in with beautiful decks.
Polished metrics, clean timelines, and thorough feature breakdowns.
And still walk out without the decision they needed.

Because they described the product. Not the experience.
There's a difference. And it's everything.

Describing the product sounds like this:
"We redesigned the onboarding flow and reduced steps from seven to four."

Describing the experience sounds like this:
"A new user can now reach their first 'aha moment' in under two minutes, without needing to contact support."

Same work, but a completely different story.
One tells leadership what you did.
The other makes them feel why it matters.

Senior stakeholders don't sit in meetings to learn about your product. They sit in meetings to decide what's worth betting on next.

And nobody bets on features.

They bet on outcomes, and on feeling like something real is going to change for someone.

When I was leading teams at Google, the PMs who got resources weren't always the ones with the strongest data.
They were the ones who could make leadership feel the weight of the problem, and the relief of the solution.

If your meetings keep turning into Q&A sessions instead of decisions, don't go back and fix your slides.

Fix your story. Then build everything you say around that.

P.S. I designed the live "Persuading People: The PM's Edge in the Age of AI" cohort to help you communicate decisions so clearly that leadership trusts your judgment, so you get seen and rewarded. More than 100 PMs have already used it to accelerate their growth. 

Register now: https://lnkd.in/d7STtuG8

The most engaged I've ever seen a leadership team during a PM presentation was when the PM said, "There was one thing we got completely wrong, and I'll get to it in a moment."

The room that had been half-checking phones suddenly wasn't.

At Google and Meta, I've spent over a decade watching PMs present to leadership.

The ones who lose the room almost never lose it because their content is weak.

They lose it because they assume that being given the floor means they have people's attention.

Those are two completely different things.

Getting 30 minutes on a leadership calendar is not the same as getting 30 minutes of genuine focus from the people in it.

Attention has to be earned, continuously, throughout the presentation, not just at the start.

The PMs who hold rooms understand this instinctively.

- They read the energy as they go.

- They know when the room is drifting, and they do something about it: a sharp question, an unexpected admission, a moment of tension that pulls people back in.

- They treat the presentation less like a delivery and more like a conversation they're actively managing.

Most PMs prepare obsessively for what they'll say and almost not at all for how they'll keep people genuinely present while they say it.

That gap is where good ideas die, not because the idea wasn't good enough, but because the room had mentally moved on before the ask even landed.

Your next presentation will be judged not just on the quality of your thinking, but on how well you held the room's attention long enough for that thinking to actually land.

Hit repost if you agree. 

P.S. If this resonates, and you want to learn how to hold a room register for my upcoming cohort, Persuading People: The PM's Edge in the Age of AI.

150+ PMs have used it to sharpen how they frame decisions, lead rooms, and earn bigger bets.

Register here: https://lnkd.in/dB7xfrvr

Something interesting is happening quietly in Product.


The PMs who feel the AI shift coming are not panicking anymore.
They’re upgrading.

Over the last few months, I’ve watched more senior PMs deliberately invest in one specific skill.

They’re sharpening how they think and communicate.

Because AI is now commendable at generating:
Strategy drafts, PRDs, competitive analyses, deck outlines, and executive summaries. In seconds.

So if your edge was “I write good docs” or “I pull great data,” that advantage is compressing fast.

The PMs who see this clearly are doing something different.

They’re building the one skill AI can’t replicate: Conviction in a room.

The ability to:

- Frame a problem so leadership feels it
- Anticipate objections before they surface
- Guide a debate without escalating it
- Turn ambiguity into a call

That skill compounds.

AI isn’t replacing PMs.

It’s compressing the average ones.

If you wait for your performance review to tell you this matters, you’re late.
If you invest before the pressure forces you to, you’re ahead.
This is why so many PMs are choosing to build persuasive clarity now.

Because it’s becoming visible who can move a room, and who can’t.

AI won’t replace PMs.
But it will compress the gap between average and excellent.

If you’re serious about being on the right side of that gap, that’s exactly why I created “Persuading People: The PM’s Edge in the Age of AI.”

More than 140 PMs have already used it to sharpen how they frame, defend, and drive decisions.

If you’re ready to be the PM leadership trusts with the biggest calls, register here: https://lnkd.in/dB7xfrvr

Next Monday will look exactly like this Monday.

Same meetings. Same dynamics. Same frustrations.

Unless you do something different.

Most PMs don’t plateau because they lack talent.
They plateau because they keep waiting for the “right time” to build the skill that actually changes their trajectory.

You already have the ideas.
The strategy.
The analysis.

But if you can’t:
↳ Frame trade-offs so leaders trust your judgment
↳ Make a skeptical room feel clear instead of confused
↳ Turn messy inputs into a confident call

...you stay the “reliable one”.

Trusted to execute.
Not trusted to lead.

That’s the difference persuasion makes.
And it’s exactly what I teach inside my live Maven cohort → Persuading People: The PM’s Edge in the Age of AI.

This isn’t about presenting better slides.

It’s about becoming the PM who can:
🔥 Rally a room around a clear direction
🔥 Influence senior leaders when stakes are high
🔥 Shape strategy through narrative, not just analysis

Enrollment closes in 3 days.

If you’re thinking, “Now isn’t perfect timing” – remember:
Next Monday will look exactly like this Monday unless you change something.

And to remove the risk entirely:
↳ 100% tuition-back guarantee
↳ Move to a later cohort if your schedule changes

The only real cost is staying exactly where you are.

👉 Enroll here: https://lnkd.in/ekgNwsyc

No alternative text description for this image

Hard truth:
Waiting feels safe.
It’s not.

The cost isn’t the tuition.
It’s another quarter being the PM who executes instead of influences.

The PMs who grow fastest aren’t better.
They just recognize moments where waiting has a cost.

This is one of those moments.

🚨 Doors close tomorrow.

Link to enroll is in the post below.

shubham srivastava

At Amazon, I have spearheaded the transformation of data architecture, establishing a robust, data-driven culture that supports strategic decision-making at scale. By architecting and deploying cloud-native AWS data lake solutions, I reduced data processing times by 30% and improved query performance by 45%, significantly advancing analytics capabilities across the organization. Over the past five years, I have specialized in designing and optimizing real-time and batch data pipelines, achieving an exceptional 99.99% data availability that ensures seamless analytics for global operations.

My leadership has been instrumental in deploying self-service analytics tools, driving a 25% increase in user adoption among over 1,200 business users. These efforts empower stakeholders to derive actionable insights, streamlining decision-making processes independently. Beyond analytics, I have developed Python-based CI/CD frameworks that automate workflows, reduce deployment time, and elevate engineering standards.

I am deeply committed to operational excellence, consistently improving processes, and fostering innovation. By aligning strategic objectives with cutting-edge data solutions, I have enhanced productivity and ensured that the infrastructure scales efficiently to meet Amazon's growing demands. My ability to translate complex technical challenges into impactful solutions has driven measurable success and empowered teams with reliable, high-performance data systems.

Senior Data Engineer

Shubham Srivastava

Right now, junior data engineers are building Spark pipelines they have never debugged, writing Airflow DAGs they do not understand, and shipping dbt models copied and pasted from AI tools.

And calling it productivity.

I get it. AI makes you fast. It removes friction. It fills the blank screen with something that looks right.

But in data engineering, "looks right" is the most dangerous place to be.

A bad API fails loudly. A bad pipeline fails quietly. It delivers wrong numbers to the right people, at scale, for weeks before anyone notices.

By the time a stakeholder catches it, three downstream teams have already made decisions on data that was lying to them.

That gap between "the job ran" and "the data is correct" is exactly where fundamentals live.

Knowing why a skewed partition turns a 2-hour Spark job into a 14-hour one.

Knowing what idempotency actually means when your pipeline retries at 3 AM.

Knowing why your watermark logic produces slightly wrong aggregations every Monday after a weekend traffic drop.

You do not learn these by prompting alonw. You learn them by breaking things, debugging them late at night, and slowly building the instinct to see failures before they happen.

AI cannot give you that instinct. It can only accelerate the instinct you already have.

I use AI every single day. For boilerplate, test case generation, and exploring unfamiliar APIs. It makes me faster in good, measurable ways.

But I use it on top of 13 years of understanding what actually breaks in production data systems. When you use AI before building that foundation, you get fast at assembling things you do not understand. That is not a productivity gain. That is technical debt with a confident face.

Master the fundamentals first.

Understand partitioning, joins, idempotency, state management, and data contracts well enough to explain every decision without opening a chat window.

Then use AI to go 10x faster. Skipping that order does not make you a faster data engineer.

It makes you a faster liability.

Principal Data Engineer @ Amazon | Data Engineering

Two Data Engineers interviewed at FAANG for a Data Engineering role.


One got rejected.
One got hired.
Same interviews.

Different grasp of fundamental data concepts.

Meta tests foundations because their stack is proprietary.

You can't learn their tools before joining. But you can master the core principles that translate to any data system.

Here's the fundamental Data Engineering stack you need to master, no matter the company you're aiming for.

Layer 1: Data Modeling & Schema Design

The foundation everything builds on.

- Normalization vs denormalization tradeoffs.
- Star and snowflake schemas.
- Slowly changing dimensions.
- Partitioning and bucketing strategies.

Poor modeling? Your queries will never scale.

Layer 2: SQL & Query Optimization

Your primary language for data.

- Complex joins and window functions.
- Query execution plans and indexes.
- Subquery vs CTE performance.
- Aggregation optimization techniques.

Can't write efficient SQL? You won't pass the technical.

Layer 3: Distributed Systems Fundamentals

How data systems actually work at scale.

- CAP theorem and consistency models.
- Partitioning and replication strategies.
- Distributed query processing.
- Fault tolerance and recovery.

Miss these concepts? You can't reason about production issues.

Layer 4: Data Pipeline Architecture

Moving data reliably at scale.

- Batch vs streaming tradeoffs.
- Idempotency and exactly-once processing.
- Backfill strategies and data quality.
- Orchestration and dependency management.

Bad pipelines? Data teams lose trust in your work.

Layer 5: Storage Systems & Formats

Where and how you store matters.

- Row vs columnar storage tradeoffs.
- Parquet, ORC, Avro characteristics.
- Data lake vs warehouse patterns.
- Compression and encoding strategies.

Wrong storage choices kill query performance.

Layer 6: Data Quality & Observability

Production data is messy.

- Schema validation and evolution.
- Data lineage and impact analysis.
- Monitoring pipeline health.
- SLA definition and alerting.

No observability? You're flying blind in production.

Layer 7: Performance & Scalability

The difference between junior and senior.

- Understanding data skew and hotspots.
- Memory vs disk tradeoffs.
- Caching strategies and materialization.
- Cost optimization techniques.

Can't optimize? Your pipelines won't survive scale.

In a world full of 10x Engineers who can't even show up to a meeting on time, I’ll take the "Reliable" engineer every single day. Talent is common. Consistency is a superpower.

Early in my career, I was lucky. I worked with leaders who did not need slides or speeches to teach ownership.

They came on time.
They followed through on what they promised.
They closed the loop, even on small things that were easy to ignore.

No one called it a "leadership framework."

It was just how they lived.

Around the same time, I was packing my bags to leave my small apartment in Bangalore for my first assignment in the US.

My landlord was this quiet uncle type. On my last day, I expected a conversation about deposit, cleaning, bills.

Instead, he handed me a short note. It was about how, for two years, I had never missed a due date, never made him chase, and always informed him if I was running late.

One line in that note, I still remember to this day: "You will do well in life because you respect other people’s time."

It also connected a pattern I had been seeing.
My managers trusted me first for small things

– joining calls on time
– sending updates when I said I would
– owning bugs instead of blaming others

Before they trusted me with big things

– owning services
– mentoring people
– eading projects that had real risk attached

Since then, I have been intentional about a few simple habits:

– I aim to be 1 to 2 minutes early to meetings, not 5 minutes late with a "sorry, got pulled into something."

– I gather context before calls so I am not hearing everything for the first time on the call.

– If I am stretched, I say it early and clearly instead of ghosting or silently slipping deadlines.

– If I take something up, I own it end to end, including the boring follow ups and cleanup.

None of this is glamorous. You cannot put "replies to messages on time" in your LinkedIn headline.

Today I mentor multiple people at work, do design reviews, and handle ongoing 1:1s.

People know that if they are stuck and they message me, I will either help or tell them when I can.

Not because I am the smartest person in the room.

Because I am predictable.

My personal rule is simple: If I take it up, I carry it fully.

This mindset also spills outside work.

It is how I try to show up for my family.

Being there for school events.

Keeping promises I make to my kids.

Showing up for friends when things are hard, not only when things are fun.

Being available is underrated.

Being reliable is remembered.

The longer I work in tech, the more I believe this:

Your stack will change.

Your title will change.

Your location and company will change.

What follows you everywhere is your pattern of showing up.

Not every blog post is worth your bookmark.
Not every Udemy course is worth your weekend.
Not every YouTube tutorial is worth your evening.
Not every LinkedIn post is worth your scroll.
Not every certification is worth your three months.
Not every new tool is worth learning before you have mastered the last one.

Information overload is the silent killer of data engineering careers right now.

I say this as someone who has lived on both sides of it.

I am a father of two. I am a husband. I am a Principal Data Engineer. My time is not infinite. Yours is not either.

And I see junior DEs doing this constantly: signing up for five courses, following thirty newsletters, bookmarking a hundred resources, and then lying awake anxious about how far behind they are.

You cannot row ten boats at once.
Pick one. Get somewhere on it. Then pick the next.

Here is what intentional learning actually looks like:

Figure out where you are today. Pick one skill that moves you two steps forward from that exact spot. Find the one resource that teaches it best. Finish it. Apply it. Then repeat.

That is it. That is the whole system.

You do not need to know Kafka, Flink, dbt, Iceberg, Spark, Airflow, and vector databases all at once. Nobody is hired for their ability to consume content. They were hired for their ability to build things that work.

The market is noisy.
The anxiety is real. I understand it.
But more resources are not the answer.
More focus is.
Consume less.

Build more. Move intentionally.

A Senior Data Engineer candidate was asked to design a feature store during his interview at DoorDash.

Another candidate in a different loop at Stripe got the same prompt.

Feature stores look simple until you add one layer of reality:

– Add point-in-time correctness? Now you need temporal joins, event-time alignment, and label leakage prevention.
– Add training vs serving skew? Now your batch precompute and online retrieval must produce identical results.
– Add feature versioning? Now models trained on v1 are in production, but v2 is live and breaking inference.
– Add backfills? Now you're recomputing 2 years of features without breaking production serving.
– Add low-latency serving? Now you're managing Redis, DynamoDB, and sub-10ms p99 latency requirements.

Here's my checklist of 15 things you must get right when building a feature store:

1. Start with your serving requirements: batch, streaming, or real-time
→ Batch = training only, precomputed features stored in warehouse. Streaming = near real-time via Kafka/Flink. Real-time = sub-10ms lookups for online inference. Each adds a layer of complexity.

2. Define your feature taxonomy: entity, timestamp, value
→ Every feature has an entity (user_id, product_id), a timestamp (when the feature was valid), and a value. This triple is your foundational data model.

3. Build point-in-time correctness into every feature
→ Training data must reflect what the model would have seen at prediction time. If you join features from the future (label leakage), your model performs great offline and fails in production.

4. Implement temporal joins that respect event-time semantics
→ When joining features to labels, use "as-of" joins: for each label at time T, fetch the most recent feature value before T. Standard SQL joins will leak future data.

5. Design for training-serving skew elimination
→ Training features come from batch jobs (Spark). Serving features come from online stores (Redis). If the computation differs, your model sees different inputs at train vs serve time. Feature store must guarantee identical logic.

6. Use a dual-write pattern for batch and online stores
→ Compute features once (in Spark or Flink), then write to both: data warehouse for training, key-value store for serving. Avoid recomputing features in two different languages (Spark + Python).

7. Implement feature versioning from day one
→ Features evolve. When you change a feature definition, models trained on v1 must still serve with v1 features. Store features with version tags: user_purchases_v1, user_purchases_v2.

8. Build a feature registry with metadata and lineage
→ Every feature needs: name, description, owner, type, version, source tables, transformation logic, freshness SLA. Without this, features become tribal knowledge.

Continued in comments

After almost 9 years at Amazon, I'll tell you what separates good pipelines from great ones.

It's never the tool.
It's always the discipline.

100 tips to write Clean Data Pipelines👇

1. Avoid NULLs in join keys
2. Sample data > mocks
3. Keep pipelines simple
4. Use DataGrip or dbt Cloud
5. Table names > inline comments
6. Split DAGs, don't monolith
7. Write fast data quality tests
8. Use strong column names
9. Schemas must fit contracts
10. Minimize SQL comments
11. Delete unused pipelines
12. Keep pipeline stages cohesive
13. Test data quality early & often
14. Master your query editor shortcuts
15. Set max SQL line width
16. Remove noise columns
17. Avoid hardcoded thresholds
18. Avoid hardcoded table names
19. Use SQL auto-formatters
20. Avoid monolithic DAGs
21. Commit pipeline changes early & often
22. Working pipeline ≠ clean pipeline
23. Comments explain business logic
24. Prefix boolean columns (is_, has_)
25. Use searchable column names
26. Don't duplicate transformation logic
27. Avoid bloated WHERE clauses
28. One transformation per CTE
29. Use consistent naming across the warehouse
30. No essays inside SQL comments
31. Link pipeline changes to tickets
32. Keep DAG inputs/outputs minimal
33. Avoid hardcoded global configs
34. Capture business logic in dbt models
35. Write repeatable data quality tests
36. Refactor pipelines early & often
37. Produce thorough data tests
38. Delete dead pipeline stages
39. Depend on table contracts, not raw sources
40. Use pronounceable table names
41. Keep proper SQL indentation
42. Write independent data quality checks
43. Don't abbreviate column names
44. Max 1 transformation per CTE
45. Use parameterized pipeline runs
46. Decouple ingestion from transformation
47. No horizontal SQL alignment
48. Use Arrange-Act-Assert in pipeline tests
49. Readable SQL > clever SQL
50. Limit pipeline task parameters
51. Use meaningful sample datasets
52. Readable pipeline > clever pipeline
53. Avoid boolean task flags
54. Hard-to-test pipeline = bad design
55. Use consistent SQL formatting standards
56. No transformation logic inside data tests
57. One responsibility per DAG
58. Write meaningful pipeline commit messages
59. Write deterministic data quality checks
60. Hide irrelevant columns in test fixtures
61. Use domain-based folder structure
62. Document your data contracts
63. Use nouns for table names
64. Review SQL before it hits production
65. Use consistent business terminology
66. Avoid storing everything as VARCHAR
67. Modular models > monolithic queries
68. Avoid pipelines with too many config params
69. Name tests: when_nulls_then_fail
70. One assertion per data test
71. Name tests after what they validate
72. Tests should fail loud, not silently
73. Never name columns is_not_deleted
74. is_active, has_churned, was_refunded
75. Assert row counts, not just "no error"
76. One output table per pipeline stage
77. Review data models with your team

Continued in comments

Here’s how I would handle the "can you pull this data for me real quick?" request as a Data Engineer. (It works 9/10 times)

I would never just say, "Sure, give me 5 minutes."

Because ad-hoc data requests turn into recurring weekly asks, and suddenly you're a human API.

Here's what I’d say instead:

"Okay! Just so I can help you long-term, is this a one-time question or something you'll need regularly? If it's recurring, I can build you a dashboard or teach you how to pull it yourself."

This does three things:

1. Forces them to think about whether this is a pattern or an exception

2. Reveals whether I should solve it with self-service or automation

3. Gives me a chance to break the "just ping the data engineer" habit

If it's truly one-time, I pull it quickly and move on.

If it's recurring, I go into enablement mode:

The key here? Scale yourself, don't become a bottleneck.

I offer: "I can pull this now, but if you need it weekly, let me show you the query so you can run it yourself. Or I can add it to the dashboard and you'll have it auto-refreshed every day. Which works better?"

This builds capability instead of dependency.

And sometimes?

They realize 6 other people need the same data, and now I've justified building proper self-service instead of answering Slack messages all day.

P.S. How do you handle ad-hoc data requests without becoming a bottleneck? Share your strategy below.

→ A senior data engineer is not “senior” because they work 12 hours a day.
→ A lead is not a lead because they answer every Slack message at 11 PM.
→ A principal isn’t a principal because they know every function in Pandas.
→ A staff engineer isn’t staff because they can run the biggest SQL queries.

In the last 7+ years, I’ve met hundreds of people just starting their careers in Data.
And honestly?

80% don’t understand what separates a junior from a senior, or why some data engineers get promoted and others get stuck for years.

Promotions and titles aren’t handed out for effort alone.
There’s a whole mindset shift that happens behind the scenes.

Let me break it down with a simple example:

Imagine your team goes on a picnic, and suddenly, you find out:
the food you brought is missing.

What happens next?

→ The junior data engineer waits for someone to tell them what to do.
 Maybe they start looking for the nearest shop if given clear instructions.

→ The mid-level quickly checks Google Maps, finds a grocery store 2 km away, and offers to go buy supplies.

→ The senior realizes there are dietary restrictions in the team.
 They talk to everyone, gather requirements, and split up tasks to get the right food, faster.

→ The staff engineer steps back and asks: “Why did we only pack food from home? Is there a better way?”
They negotiate with a nearby restaurant for a group meal and handle logistics.

→ The principal asks: “Why are we eating at all if the weather is about to turn bad?”
They move the whole picnic to a better location with indoor dining and future-proof the team’s outings.

So, next time you wonder why someone got promoted:
It’s not always about who’s working harder.
It’s about who’s thinking bigger, seeing patterns, and driving real change.

If you want to move up, start changing the way you approach problems, not just how many problems you solve.

writing the DAG
scheduling the DAG
getting paged by the DAG

If I ask a Principal Software Engineer for system design books, they can easily give me a list of 10: Grokking. Alex Xu. Designing Data Intensive Applications. A bunch of classics.

If I ask most Data Engineers the same question, the answers suddenly become vague.
A few blog posts here. A random YouTube playlist there. Maybe a conference talk or two.

The reality is:

– Most senior Data Engineering roles now have a dedicated system design or architecture round
– We are expected to talk about trade offs, SLAs, freshness, reliability, cost, lineage, quality
– Yet there is no single "gold standard" resource that does for Data Engineering what Alex Xu did for backend system design

I felt this very strongly recently.I finally got time to sit with Alex Xu’s System Design Interview. The structure, the repetition of patterns, the way he scales the same idea across different problems - it is genuinely good. It is a staple for backend engineers for a reason.

But as a Data Engineer, it also highlighted a big gap.

Designing a feed service or URL shortener is not the same as:

– Designing a Lakehouse with Iceberg or Delta
– Modeling data contracts at the source so downstream does not break silently
– Planning for backfills, late arriving data and schema evolution
– Architecting pipelines that support LLM and RAG workloads without blowing up cost

We borrow patterns from backend system design, but a lot of our pain is unique:

– Batch plus streaming together
– Multiple storage layers with different guarantees
– Orchestrators, quality checks, catalogs, contracts, governance
– Stakeholders who care more about trust in the numbers than response time in milliseconds

Which is why I keep thinking: We probably need a dedicated "Data System Design" playbook that speaks our language.

Not just a chapter in a generic book.

An actual guide that:
– Uses data problems as the base
– Shows patterns for ingestion, modeling, contracts, lineage, SLAs
– Includes examples that cover Lakehouse, modern warehouses and ML / LLM pipelines

Curious what you think….

Do we need a dedicated "System Design" guide for Modern Data Engineering?


The author can see how you vote. Learn more

taplio

The Battle

Roshan Gupta

Shubham Srivastava

Followers

Posts

Engagements

Likes

Performance ScoreRoshan Gupta's LinkedIn Performance

Performance ScoreShubham Srivastava's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Roshan Gupta

Shubham Srivastava

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Roshan Gupta's LinkedIn Performance

Performance ScoreShubham Srivastava's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreRoshan Gupta's LinkedIn Performance