Data Leadership Consultant | ex‑Meta | Helping data teams deliver more, faster.

As an Independent Data & Technology Consultant, I help businesses unlock the full potential of their data - transforming it into actionable assets that drive efficiency, innovation, and measurable value. 

Here are some ways I can help you:
- Conceptualizing and implementing business metrics end-to-end
- Modelling the data in an easy-to-use way so it can be consumed by different functions in the organisation
- Setting up data infrastructure
- Building a data ingestion layer for multiple sources
- Platform migrations
- Assessing and identifying gaps in your current data platform and analytical capabilities
- Improving the accuracy of your analytics by improving Data Quality and Governance

About me
-----------
Highly accomplished Data Professional with a decade of experience driving innovation in academia, startups, and tech giants like Meta. My journey began in Geodesy and Geoinformatics (MSc), evolving from GIS and Remote Sensing to Software and Data Engineering.

Proven track record as both Individual Contributor (IC) and Manager, leading high-performing teams and architecting scalable solutions. Passionate about harnessing data to drive impact. Committed to mentoring and growth.

At Meta, I led Data Engineering teams responsible for the backend ranking systems that power ads across all platforms. I built and scaled the London team from inception, growing it from 0 to 9 engineers. My primary focus was leveraging data to unlock incremental business value and establishing robust analytical measurement foundations.

Outside work, I cherish time with my loving family - my wife and precious daughter.

It's no secret that soft skills are the secret sauce of high-achieving individuals. This is now more true than ever, with AI pushing boundaries and unlocking technical expertise for non-tech people.

One massive differentiating factor I've observed is proactiveness. I can't stress enough how valuable this skill is. It’s a superpower regardless of your role, whether you're an IC or a Manager.

If you are leading a Data Engineering team, these are the behaviors you should be looking for:

- Anyone can spot a failing Airflow task. That's the easy part. A proactive engineer fixes it. A proactive senior engineer proposes the architectural changes to prevent the issue from appearing in the future.

- Passive folks stay silent if things are going well. But no status update means uncertainty. Proactive engineers keep everyone updated regardless of whether things are smooth or there’s a fire.

- Passive execution is just being service-oriented. "You tell—I do." Proactive DEs challenge the status quo. They keep asking why. They want to understand the business impact and the actual value of their work.

- Volunteering to 'get the shit done' vs. just being supportive. Thumbs-ups and cheerleading are great, but nothing changes unless someone takes ownership of the problem.

As technical barriers fall, soft skills are more important than ever. If your team is only waiting for tickets to come - they’re already falling behind.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

Slawomir Tulski

Data Leader | 10+ Years of Industry Expertise
---------------------------------------------
Highly accomplished Data Professional with a decade of experience driving innovation in academia, startups, and tech giants like Meta. My journey began in Geodesy and Geoinformatics (MSc), evolving from GIS and Remote Sensing to Software and Data Engineering.

Proven track record as both Individual Contributor (IC) and Manager, leading high-performing teams and architecting scalable solutions.

Currently, I lead Data Engineering teams at Meta, focusing on backend ranking systems that power ads across all platforms.

Passionate about harnessing data to drive impact. Committed to mentoring and growth.

Outside work, I cherish time with my loving family - my wife and precious daughter.

A "Senior" title is relatively easy to handle. Senior behavior are harder to find.

True seniority isn't about how long you've been using a framework. It’s about how you handle complexity, people, and business value.

If you’re wondering if your team is truly operating at a senior level, look for these 5 signals 🔍
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

How to tell if your engineers are senior… or just loud

Data Engineering for Leaders: Batch Processing - Because some things are better served in bulk.

In a world obsessed with "real-time," there is a common misconception that faster is always better. Trying to process every single event the microsecond it happens is often overkill. It’s expensive, technically complex, and usually unnecessary for high-level decision-making.

Batch Processing is the engineering fix for this.

Think of it like your household laundry. You don't run the washing machine for every single sock the moment it hits the floor; you wait until you have a full basket. It’s more efficient, saves energy, and gets the job done at the right time.

Batch processing collects data over a specific period - be it an hour, a day, or a month - and processes it all at once. It is important to note: Batch does not mean "once per day." You can still get frequent updates that satisfy all your requirements without paying the premium for "true" real-time.

Batch pipelines are easier to build, audit, and fix. For you, this translates into faster and cheaper delivery of insights.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

Data Engineering for Leaders: Data Lineage - Because "I think the numbers are right" isn't an answer.

Imagine you’re in a board meeting. The CEO points to a churn metric and asks a simple question: "Where exactly did this number come from?"

If your team’s answer involves hunting through three different departments, five Slack channels, and a "master" spreadsheet that only George knows how to use - you’ve got a problem.

In the real world, data movement is messy. It’s ingested, transformed, renamed, and aggregated. By the time a KPI hits your dashboard, it’s a "Magical Number." You see the result, but you don’t understand how it happened. But you must have hard evidence.

Data Lineage is the engineering fix for this.

Think of it as accounting for everything that happens to your data. It provides a permanent map of what, where, how, and for what purpose data moved from the source to the screen.

With proper Data Lineage, you get:
- Auditability: Prove the origin and state of any metric within an hour.
- Impact Analysis: Know exactly which reports will break before you change a source system.
- Governance & Privacy: Ensure compliance by seeing where sensitive data is stored and what is using it.

You wouldn’t trust a financial audit without a paper trail. You shouldn't trust your data without lineage. It’s the difference between magic and science.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

Honestly... if your KPIs go up, down, or do a backflip and absolutely nothing changes? Stop tracking them.

You aren’t "monitoring" anything. You’re just looking at a bunch of numbers that make everyone feel like they have control.

The Lesson: If a measurement doesn't force a decision - it’s just noise.

It’s easy to keep adding metrics because they look professional on a dashboard. But for every single measurement, you need a cold, hard answer to one question: "What do we change when this moves?"

Remember: these metrics aren't free. Every useless graph is a liability that your data engineers have to build, monitor, and maintain. When you hoard "just in case" data, you’re taxing your team’s focus for zero ROI.

Human brains don't scale to dozens of metrics on a single view. By the time you’re looking at twenty, you’ve lost the signal. You aren't informed; you're just overloaded.

Metrics and KPIs should give you clarity and allow action. Not the opposite.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

One reason I write on LinkedIn is to hold myself to a high bar. I couldn't look in the mirror if I lured you with smart-sounding advice while doing the opposite in my own career.

That’s why I’m finally launching The Dark Side of Data Engineering on Substack.

Why now? Honestly, it’s a terrible time. I’m pushing my business and settling into a new country. But I always say "you'll never be ready" and if I wait for the perfect moment, I’ll be waiting forever. I’m done with excuses.

It’s not for everyone. It’s for the group closest to me - Data Engineers. I’ll be focusing on the Individual Contributors who are already in the industry and want to go beyond the Junior and Mid-level plateau.

What to expect:
Think of this as having an online Meta DE manager in your corner. Don't expect tooling tutorials or SQL analytical functions explained. I'll be showing you what it actually takes to reach Senior, Staff, and Leadership levels - the things that matter once you've already mastered the tech and tooling.

The plan:
- A new article every 2 weeks.
- Extended LI thoughts that need more room to breathe.
- The same bad jokes and serious engineering talk you’d expect from The Dark Side of DE.

I’m figuring out the paid vs. free tiers as I go. For now, everything is free.

The first piece is live today. If you want to see more - subscribe there!

Link in the comments!

Having a lot of work doesn't necessarily mean delivering great value or doing the right things. It can also be a sign that something's going wrong. 

Think about these examples:
- Tons of ad-hoc data requests: Is it appreciation for data or lack of self-serving and foundational data assets?
- Constantly unblocking critical pipelines: Hero mode or failed to fix recurring issues?
- Busy on-call shift: just one of "those weeks" or shaky data assets that need improvement?

Unfortunately, sometimes we're caught in a vicious cycle. A lot of extra work is created due to existing gaps, but there's no time to fix those gaps because of all the extra work.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

Why we sabotage simplicity? Because simplicity is invisible.

When an engineer rescues a "clever" but fragile pipeline, they look like a hero. When they build a system so simple it never breaks, they look like they aren't doing anything.

Our brains equate effort with value. We notice the fires, but we ignore the silence of a system that just works.

If a solution feels "too easy," we assume it’s not important. We’ve conditioned ourselves to believe that complexity equals competence.

It’s a trap.

As a leader, you have to break this cycle. If you want simplicity, you have to reward it:
- Stop praising the "hustle" and start rewarding the outcome. Effort only matters if it actually moves the needle.
- Make the "fire prevention" as prestigious as the firefighting. Don't wait for a crisis to notice who is doing the best work.
- Redefine what a Senior is. Real seniority isn't about navigating complexity - it’s having the discipline to avoid it in the first place.

Stop rewarding cleverness. Start rewarding impact.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

I love the days when I'm reading about AI revolutionizing the world while my dbt sources.yml refuses to run a freaking macro.

As an experiment I ran the problem through several models (including the newest Claude) to see if they could find a creative solve. All of them failed. But... the way they failed matters a lot.

Good ones: quickly explained me why what I try to do is not possible
Bad ones: "Here is a completely made-up solution that will break everything"

This is not a small difference. We keep saying how AI makes us more performant. Now think about the time I'd waste chasing all of these made up traps...
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

“No matter how good a life you lead, you won’t please everyone. Someone will be glad to see you go.” - Marcus Aurelius

You will never satisfy everyone around you. There’s no sense in trying - sooner or later, you’ll fail.

Instead, focus your energy on over-delivering for:
- Your people (The ones doing the work)
- Your direct stakeholders (The ones moving the needle)
- Those you want to impress (The ones who open doors)

Good leadership isn’t about being liked; it’s about driving a shared vision and empowering your team. If you aren't occasionally disappointing someone, you are likely burning out yourself or your best engineers.

Want a tangible example?

Every "Yes" to a low-value, ad-hoc request from a fringe stakeholder is a silent "No" to the core infrastructure your team actually needs.

Prioritize the mission over the applause.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

If you only talk to your own team, you’re missing the big picture.

Of course, it’s easier to talk to folks from your own department. They know what you do and how Analytics works. They understand you and speak the same language.

Yet you’re missing a lot if you’re just sticking to that. You need to go broader. Talk to Business, Marketing, Sales, Engineering - any and all of the different leaders in those orgs.

This will help you:
- Broaden your perspective to understand priorities better. What problems do they have? What keeps them awake at night? Can they access the data they need? How hard is it?
- Educate them and show what great work Analytics does. Show them how Analytics can help - and why it really matters.

Talking, asking, and listening is real work. It’s not easy, and the fruits of it aren’t always as direct as in other activities. However they make a big difference over time!
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

Without real commitment from leadership, "It’s an amazing opportunity" is just code for:

"Clean up this impossible mess so I don't have to look at it. If you fail, I’ll blame your attitude."

The Lesson: Opportunity shouldn't be a trap.

Proactivity and rejecting the status quo are top-tier qualities in any manager. You want to nourish them. But that requires more than just "encouragement" - it requires a mandate for change.

In Data Engineering, this distinction is critical.

If you give a leader a mess of undocumented legacy pipelines and broken data contracts, you must give them the resource allocation and political backing to fix the architecture.

Otherwise, you aren't offering an opportunity; you're just delegating technical debt.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

You won’t notice the dystopia arriving. You’ll just wake up one day and realise it’s already here.

Boring Dystopia is a concept that originates in Mark Fisher’s thought. It’s a powerful idea that can be applied to the corporate world. 

Dystopia doesn’t come and announce itself loudly with sirens and fireworks. Its arrival isn’t broadcast on the news. You won’t even notice that it’s already crept in.

It builds slowly over time through:

(1) Gradual erosion
- A new process that’s a bit superfluous and dumb.
- Another broken report.
- The working day extended by 10 minutes here and there.
- More repetitive, mundane tasks that are quick to do.
- A system patched because the bigger change is too hard.
- An attendance list added to the meeting agenda.

(2) Alienation
- Others seem fine with how things are.
- People go about their business or get promoted.
- Maybe that’s just something I don’t understand?
- Anyway, all those problems aren’t personal - it’s just a workplace. This isn’t my problem.
- I don’t even like many of the people I work with.

(3) Normalisation
- This is how we run things here.
- It’s always been like this.
- Nothing weird about it.
- Other companies have inefficiencies too.

Day-to-day operations matter. They build over time - positively or negatively. When it grows, the momentum that’s been built either way will be hard to stop.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

Many leaders think scaling an analytics team means hiring more specialists. But Meta does the exact opposite.

There is this concept of Artisan and Factory analytics team model. Not all analytics teams are built the same. Here’s why that matters:

(1) The Artisan Model
- At a high-level, Artisan is a small team of talented generalists.
- They are able to find opportunities, plan, execute and present.
- They are autonomous, good at communication and pushing the work end-to-end.

I often hear this working model in the context of startups where small teams push the boundaries and deliver amazing things while wearing different hats.

It does not need to be coupled with the company size or maturity of the organisation though. E.g. Meta is running on this despite being a huge mature corporation.

(2) The Factory Model
- The factory model takes a different approach. This relies on specialisation.
- People are responsible only for certain narrow parts of the process.
- There are folks who gather business requirements, others model the data, yet another cleans and implements the pipeline, etc.
- There are more functions hired and more management involved.
- On the other hand it is supposed to scale better as such an approach tries to remove bottlenecks and optimise the process as in the factory line.

In reality - it’s NOT a battle between Good and Evil. Those are just different models one could apply to their organisation. They have their own pros & cons.

There are companies that function well on either of them or anything between as in reality you rarely find the purest form of implementation.

Why must Leaders care about this? 

They need to build and steer their organizations consciously. Those working models impact most of the crucial processes like for examples:
- Hiring
- Depth and width of management chain
- Team sizes
- Responsibilities
- Team culture
- Incentives
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

Your data shouldn’t feel like a money pit.

Many companies assumes their rising cloud bills are just "the cost of doing business." But more often than not, it’s not a tech problem - it’s an efficiency problem.

You don’t need to be a Lead Engineer to spot where the leaks are. You just need to know where to look. In this carousel, I’m breaking down the 4 red flags that prove you’re overpaying for data you aren’t even using.

The good news? Trimming the fat doesn't just save you money today - it makes your entire team faster and more agile for the future.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

4 Hidden Red Flags Killing Your Data Budget

The ultimate litmus test to see if your org is actually data-driven:

Present evidence-based data that contradicts a key stakeholder's gut feeling.

What happens next?
- Do they revise their decisions based on the new reality?
- Do they ask for the methodology to understand and learn?
- Do they simply ignore the data?
- Or do they discard the evidence and cherry-pick data that fits their opinion?

The Lesson: A company is not data-driven because it has many dashboards. It is when it uses data to form hypotheses, validate them, and respond to the results.

You may have hundreds of dashboards and pay millions in cloud bills, but it won’t be enough if the organization fails that one test.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

What does it actually mean to be a Data Engineer in 2026?

I’m joining DataTalksClub live to answer that question and more. I'm planning to have honest conversation about the real-world state of the industry - from the confusion surrounding roles to the impact of AI on our ecosystem.

Don’t miss it - join us and bring your questions!

Our next live podcast episode is right around the corner! Join us, as we chat with Slawomir Tulski about Data Engineer Career in 2026: Roles, Specializations, and What Companies Look for. You won't want to miss this one!

He’ll cover: 
🔹Why does the term Data Engineer still create confusion across the industry
🔹​The different specializations inside Data Engineering and how they differ in practice
🔹​How the same role changes depending on company maturity and team structure
🔹​What companies look for when hiring Data Engineers
🔹​Common career traps and misleading job descriptions that engineers should watch for
🔹​How the role may evolve as AI transforms the data ecosystem​

Register here: https://luma.com/ukf4tr7l

Data Engineering for Leaders: Technical Debt - Because winning today shouldn't mean failing tomorrow.

Optimizing for the short vs. long term is not only a challenge in finance; this tension is at the heart of software development.

The concept is simple: you (rightly) shouldn't build "perfect" software, fix every sub-optimal line of code, or migrate every legacy platform. If you try, you’ll lose your agility and stop delivering new value entirely.

However, every tradeoff, quick patch, and "temporary" solution adds up to create Technical Debt. Just like financial debt, it eventually starts to drain you:
- There are higher engineering maintenance costs (like interest rates you need to pay).
- Delivery velocity goes steadily down (you can't buy new things because you pay too much for your debts).
- It increases the risk for any parallel project (you may be forced to drop every new investment to take care of debt; otherwise, you're insolvent).

In Data Engineering, the symptoms are specific and painful:
- Simple requests will take a long time to fulfill. You can't rely on data for decision-making because insights are always lagging too much.
- There will be many half-broken or conflicting dashboards, leading to constant confusion and discussions.
- Your reports will be late and inaccurate. This creates a lot of frustration and a lack of trust.

Despite the above, the costs of running the team will remain very high.

But the analogy doesn't end there. Just like in business, not all debt is bad. It’s a tool for speed. The goal isn’t to have zero debt - it’s to manage it wisely so it doesn't manage you.
—-—
💾 Save this so you don't have to go looking for it later. 
♻️ Repost if you think more people should see this.

For months now, I’ve committed to pairing my daily thoughts with GenAI images. Every single day, I’ve included a graphic just to see if it makes a difference for reach.

The reality? I’m spending 10+ minutes per day making sure the image doesn't look like 💩. That’s an hour a week I could spend writing better content instead.

Looking at the numbers... I don’t see any difference. But screw the numbers - it’s my audience I care about.

Does the image actually make you stop and read, or is the text enough? Do you find them engaging? Informative? Or maybe graphics is a good idea but just my stylistics puts you off? I'd love your unfiltered take.

I’ll still keep visuals where they’re actually necessary, like:
- Weekly Memes (with a lesson, of course).
- Data Engineering for Leaders (the notebook-style concept explains).
- Weekly Carousels.

But for the rest? I need to know if the effort (and the compute used to generate this stuff) is actually worth it.

Let me know in the poll!

taplio

Slawomir Tulski's Linkedin Analytics

Slawomir Tulski

Performance ScoreSlawomir Tulski's LinkedIn Performance

What is Slawomir talking about?

Slawomir Tulski's Best Posts (last 30 days)

Ready to grow your LinkedIn presence?

taplio

.css-6ytdur{background-image:linear-gradient(284deg, rgba(39, 43, 72, 1) 0%, rgba(33, 47, 183, 1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Slawomir Tulski's Linkedin Analytics

Slawomir Tulski

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);}Slawomir Tulski's LinkedIn Performance

What is Slawomir talking about?

Slawomir Tulski's Best Posts .css-1emoo84{font-size:var(--chakra-fontSizes-md);font-weight:400;color:var(--chakra-colors-gray-600);}(last 30 days)

Ready to grow your LinkedIn presence?

Slawomir Tulski's Linkedin Analytics

Performance ScoreSlawomir Tulski's LinkedIn Performance

Slawomir Tulski's Best Posts (last 30 days)