Amanda Zhu

Co-Founder

Recall.ai

✔︎ Who gets results using Recall
∙ Product teams at fast-growing startups who want to go to market in days instead of months. 
∙ Engineering teams that don't want to burn engineering hours maintaining their meeting bot integrations.

✔︎ Here’s what we do
We provide one API for meeting bots on every video conferencing platform. We abstract over differences between meeting platforms (even for platforms that don’t have an APIs). We run, monitor, and scale the complex real-time video infrastructure.

Interested in learning more? Go to: https://recall.ai 🌐

Perfect Recall (W20)

Grew to thousands of users, signing on annual deals with customers like Faire and Quora

Edisoft Inc

Software Engineer

❗Your engineers are wasting 6+ months building meeting integrations.❗

Why? Because every platform (Zoom, Google Meet, MS Teams) has different APIs, constant updates, and hidden edge cases.

❌ Building in-house = wasted engineering time.
❌ Maintaining it = ongoing distractions from core product development.
❌ Scaling it = reliability nightmares that break at the worst time.

That’s why we built Recall.ai.

✅ One API to handle every meeting integration for you.
✅ Ship in days, not months.
✅ Powering 1200+ companies, from early-stage startups to billion-dollar SaaS teams.

🚀 See how Recall.ai can save your engineers months of work

Amanda Zhu (Hiring SEO Manager)

YC W20 Demo Day: Our batch mates raised $5M, $10M, $20M.
We raised $300K.

We went home with a handful of meetings and a lot of “maybe later.”
A few people told us to give up.
Some said the market was too small.
A few more said, “If Demo Day didn’t pop, it’s probably not going to work.”

For a long time, it felt like they were right.

We spent the next five years building.
No big announcements.
No attention.
Just customers, product, and the same question every year:

“Are we still alive because we’re onto something… or because we’re too stubborn to quit?”

This year, when we announced our Series B, we finally used the LaunchHN we never used at Demo Day.
It felt like closing a loop we opened five years ago.

Demo Day results don’t predict outcomes. The work after does.
We’re grateful for YC and want to support other YC companies however we can.

If you’re YC (any batch), drop a comment with what you’re building. Would love to connect and help where we can.

The API for meeting recording | Co-founder at Recall.ai

No alternative text description for this image

I’m COO and co-founder of a $250M company. At our company, COO doesn’t mean “operations.”

It means owning whatever the company needs most that the CEO shouldn’t.
At Recall.ai, that’s GTM.

The COO role took us through 4x growth YoY and into consistent enterprise inbound.

My co-founder and CEO, David, owns product, technical direction, and long-term vision.
I own making the machine that attracts and closes customers.

That includes:
- how we market
- how we price and package
- how sales, marketing, and self-serve actually work together

The COO role isn’t a template.
It’s a response to a constraint.

Some companies need someone to scale customer operations.
Others need someone to build a sales machine.
Others need someone to turn early traction into repeatable growth.

We needed someone to own GTM end to end.

And once that constraint moves, the job changes again.

Closed my 5th seven-figure deal last month. Finally saw the playbook everyone says doesn’t exist.

Yes, there’s no script, formula, or guaranteed sequencing for enterprise deals.
But there is a scaffold.

Every seven-figure deal eventually hits the same checkpoints:
First meeting → demo → technical validation → PoC → PoC sign-off → procurement → security → legal → contract.

But the real work isn’t the checklist.
It’s what you do when you’re stuck between checkpoints.

Here are the things that have actually moved our seven-figure deals forward:

- Backchanneling to learn who’s worried, who’s confused, and who’s blocking silently
- Flying out when a senior stakeholder joins late and wants to “reassess”
- Hosting a small dinner to align teams who weren’t looped in until mid-PoC
- Getting warm intros from investors or any mutual connections
- Running security + procurement in parallel so they don’t choke the timeline later
- Showing up early to events stakeholders are speaking at
- Going backwards a checkpoint when something collapses, instead of forcing the next step

None of these moves look like “selling.”
None of them follow a proper sequence.

That’s why people say there’s no playbook.
But there is a framework and checkpoints you have to hit.

What’s the most creative thing you’ve done to get to the next checkpoint?

Today, running a $250M company, I have 60 minute 1-1s with every person in our GTM team. This is why this works.

As the company scaled, I was tempted to compress 1-1s, but I found 30 minutes to be too short for anything real to surface.

In 60 minutes, people bring up things they wouldn’t say in hallway conversation or a weekly standup:
- the complex problem they’ve been stuck on but can’t articulate yet
- the thing they “don’t want to bother you with”
- the early signs of misalignment that only show up in long-form conversation

And once you hear those things through, 90% of the time, the solution becomes obvious because your direct report finally had room to talk through it.

The hour matters because:
- trust takes time to reveal itself
- sometimes the problem is unclear until it is said out loud
- a single good conversation can make the next two weeks dramatically more productive

It’s also a very inexpensive and effective use of the hour.
One hour every two weeks prevents months of misalignment.

The only reason I know this is because I started to skip them.
And I felt the disconnect immediately.
Problems started to compound.

This has been the highest-ROI GTM habit I’ve kept.
Even at this stage.

I sent a $1.2M contract way too early. Their CTO shut the deal down on the spot, and that’s what saved it.

The VP of Engineering was excited and wanted to move fast.

No demo.
No deep-dive.

His team validated the technical fit in a day and told him, “We can run this in production.”

So after the disco call, we kicked off everything in parallel:
- Security
- Procurement
- Legal
- Commercials

Five separate threads. 
None of them talking to each other.

On Day 5, the VP asked for the contract.
I sent it.

Two hours later, the CTO rejected the contract and replied:
“Do not move forward.”

The email exposed the problem:

Finance hadn’t synced with Eng team.
The CTO had no context at all.

The urgency was still there, but the contract exposed the lack of alignment.

It forced every stakeholder to raise their hand.
It gave us the full buying committee in one shot.

Once everyone surfaced, the work was straightforward:
security review, pricing, fallback plan, legal terms, technical requirements.

The contract kicked off what I should have done from the start.
Coordinate.

The deal closed on Day 47.

In enterprise sales, a “no” isn’t a rejection.
It’s the moment the deal finally starts.

Here’s how I got my first 100 customers using Google Search and a Word doc.


2021. No fancy tools.

The hardest part wasn’t writing the email.
It was knowing who to send it to.

I started with one simple question:
“Who needs this now?”

Not “who might be interested.”
Not “who fits a persona.”
Literally: “Who records a lot of meetings?”

That gave me the first 10 names.
Those turned into a seed list.
Then I Googled their competitors.
Then their lookalikes.
Then anyone launching adjacent products.

My entire list came from:
- “{Company} alternatives”
- “Tools like {Company}”
- “Best {category} software”
- “Companies hiring {ICP role}”

Every name went into the Word doc.
Every name got outreach.

That’s how I got our first 100 customers.
And it’s even easier now.

There are tools that can make this faster today:
- Clay 
- Apollo.io 
- ZoomInfo 
- Lusha 
- BirdDog 

But it all still comes down to the same thing:

If you don’t know who needs your product, no tool will save you.
If you do know, even Google Search is enough.

The list is everything. The outreach is just execution.

Who am I missing? What tools or strategies do you use to build your seed list?

If you’re not a YC founder, give up.


If you didn’t IPO before kindergarten, give up.
If you didn’t vibe code a chatgpt wrapper in the womb, give up.

It’s satire, but the anxiety underneath it is real.

Truth is, there is always a valid reason to not start.

The odds will always be stacked against you: too old, too young, too late, too competitive, too uncertain.

And those justifications are all true - at least statistically.

But if everyone listened to the statistics,
There would be no exceptions.

Every breakthrough we celebrate today comes from someone who started despite the numbers, not because of them.

So it might’ve been better if you started earlier.
But you’re not too late to start now.

We 3x’d revenue, raised a Series B, shipped a new product, and launched self-serve. And somehow… it’s not even the end of the year.

This was a rewiring year.
Three things hit at once:

-----

1/ Product inflection: Desktop Recording SDK expanded the entire surface area.

We had to rewrite how we:
- pitch the product
- allocate engineering time
- talk about Recall.ai in one sentence

Multi-product changed how we describe the company.

-----

2/ GTM inflection: self-serve forced us to rebuild the funnel.

Pricing, onboarding, docs, support load - everything we assumed was “done” had to be rethought.

The questions changed too:
“Is the pitch working?” became “Which product is the pitch for?”

-----

3/ Sales inflection: larger customers changed the motion.

The deals went from:
- one stakeholder → buying committees
- simple technical validation → multi-threaded reviews
- one product fit → cross-product conversations

We had to learn a new type of selling.

-----

The story of this was recognizing that every new product, every new customer segment, and every new GTM motion changes the operating model.

It’s a good reminder that companies don’t grow in straight lines.

We are hiring cracked engineers in SF to build foundational infrastructure for the world’s largest dataset. We ran 18 million EC2 instances last month alone. Running highly-optimized real-time Rust processing 3TB/second of raw video.

If you are cracked please DM me with your GitHub! We have a rigorous candidate review process so anyone who gets an interview with us will get a free bundle of the seminal K&R’s C, TLPI, TCP/IP Illustrated Vol 1 (some staples from around the office).

Three seven-figure deals hit at the same time last quarter. That was the first week I couldn’t keep up.

When it was one big deal, I could brute force it.
Know every stakeholder. Read every email. Maintain perfect context.

But when three landed at once…
My AE had to bring me up to speed.

“Here’s where Deal A is blocked.”
“Here’s what Deal B needs from you.”
”Here’s the one meeting you must take for Deal C.”

I wasn’t running the deals anymore.
I was being deployed.
And I had to get comfortable with that.

Because in seven-figure territory, the AE has 10x more context than the founder:
- They’ve been on 20+ calls you haven’t.
- They’ve built the relationships.
- They see the entire chessboard.

I get parachuted in for one conversation that only I can unblock.
Not to quarterback the whole thing.

The shift was uncomfortable.
Before: I managed the deal, and the team executed my plan.
Now: The team manages the deal, and I execute their plan.

Your instinct is to grab the wheel.
To be “the closer.”
To jump into every thread.

But if you try to personally run three seven-figure deals at once, you will burn yourself and the deals to the ground.

I’m still figuring this out.
Where do you draw the line between stepping in and stepping back?

157 YC companies launched last week. We went through YC W20 and know how fast the spotlight moves. These 25 deserve a longer look.

item (YC F25) : the AI-native CRM that works for you

Fixpoint : Uber for AI trainers and human-in-the-loop supervisors

Bear (YC F25) : marketing stack for AI agents

Zarna (YC F25) : AI associates for private capital

Clicks (YC F25) : Computer use agents to automate all back-office work

Claybird (YC F25) : the full-stack AI ad company

Bluma (YC F25) : the new media distribution company

The Hog (YC F25) : AI demand gen for companies

Mantle : build internal agents to automate back office work with one prompt

Karumi (YC F25) : 24/7 AI demo agent for B2B SaaS

Denki (YC F25) : the full-stack AI financial audit firm

Absurd (YC F25) : AI brand and performance ads at scale

Caddy : the voice interface that works proactively for you

Crunched : Excel AI analyst for power users

HireGlide (YC F25) : AI screening and matching for hiring teams

Aside (YC F25) : live answer copilot for enterprise tech sales

Imagine AI (YC F25) : your AI clone for B2B content

Leadbay (YC F25) : discover and qualify leads that couldn’t be found before

Lexi (YC F25) : AI associates for corporate law

Logical (YC F25) : a proactive desktop copilot (Clippy but actually good)

Sciloop (YC F25) : AI co-scientist that automates ML experimentation and analysis

Cranston AI (YC F25) : full stack AI accounting firm

Questom (YC F25) : AI agents for B2B inbound sales

Sorce (YC F25) : Tinder for jobs

Uplane (YC F25) : replace marketing agencies with AI

p.s. to make it 26 ComplyDo (YC F25) : global compliance for enterprises

We get 6k signups every month, and 100% of our pipeline is inbound. But our inbound didn’t start with marketing.

For the first year, everything was founder-led outbound.

No paid acquisition.
No growth team.
No scalable channels.

Just me emailing every team building anything with meetings in it, trying to figure out who exactly needed this infrastructure the most.

Our first dozen customers came in one by one.
And every single one of them shipped fast.

One day a prospect booked a demo and said:
“We asked [Company Name] how they built their meeting recorder. They told us: Recall.ai.”

Then it happened again.
And again.
Different customers.
Same story.

We weren’t doing anything different.
Our customers were.

That’s when I realized inbound comes from density.

For us, it was enough teams in the wild using our product that other builders would bump into and ask how they did it.

Once that happened, inbound went from a trickle to a stream.

Outbound gave us our first customers.
Those customers gave us inbound.

Hired an AE who spent 6 years doing animal husbandry in a therapeutic research lab. Weird background for sales. Perfect background for enterprise.

It’s easy to filter for linear pedigrees: SDR → AE → Logos.
But seven-figure deals aren’t linear.

During the interview, I asked what that early chapter taught her.

She didn’t talk about grit, or passion, or some polished framework.

She said:
“You learn quickly that things don’t move on your timeline.
You show up every day, observe carefully, document what’s changing, and adjust.”

That sentence told me everything I needed to know.

Enterprise sales is filled with:
- periods of silence
- subtle shifts in internal politics
- a champion who loses influence
- a process no one warned you about
- an evaluator who stops replying and won’t say why

You win these deals with pattern-recognition, patience, and a calm emotional baseline.

Her background was evidence she had already lived the part of enterprise sales you can’t teach.

When you’re hiring for high-judgement roles, don’t over-index on linear experience.
Look for the work that shaped how they think.

Oh… and we’re hiring for Strategic Account Executives.

I’ll say the quiet part out loud. 
I landed our first seven-figure deal because someone was willing to stick their neck out for me.

Not because we had SOC2.
Not because we looked like an enterprise vendor.
But because one senior leader said, “I’ll take the hit if this goes wrong.”

What I didn’t understand at the time was that they already liked our product, and were evaluating the risk of working with us.

Could we support them?
Will we be alive in two years?
Would choosing us blow back on them?

And the wild part is that none of these questions get answered in one meeting.

Mid-market = 1 person making a decision.
Enterprise = 15-50 people trying to agree on anything.

Half the calls I had with them were their own teams catching each other up.

I wasn’t selling.
I was project managing a group that didn’t work for me.

Our first enterprise deal actually died twice.
Both times, our champion dragged it back to life.

That was the moment I finally understood who actually closes these deals.

Not the founder.
Not the AE.
The person inside who’s willing to risk their reputation on you.

Funny thing is… every founder I know has a completely different story about how their first enterprise deal actually closed.

My first 97 discovery calls all failed for the same reason.
I didn’t know what I was actually supposed to accomplish in the discovery call.

I knew the end goal was to close the deal.
But I didn’t understand the milestones I had to hit on the way there,
which made every call awkward.

Here’s the core insight:
Discovery → Demo → Trial → Close.
This is the simplest version of a sales process framework.

The goal of discovery is simple: if the prospect is qualified, the next step is to get them on a demo.

In discovery, you want to understand them deeply: use case, timeline, constraints, and decision criteria.
In the demo, they start to understand your solution more deeply.

Once I shifted my disco call goal from “sign the contract” to “book the demo inside the call,”
Everything clicked.

In sales, the goal of every meeting is to book the next one.

This is episode 1/9 of what good discovery calls actually look like.
Stay tuned for part 2: the one question that sets the tone

Adobe overpaid for Semrush.
$12/share in cash for what was trading at $6.89.

Everyone’s talking about the future of SEO vs GEO, but no one’s asking why Adobe paid a premium to buy Semrush now.

1/ Adobe bought Semrush above-market value at $1.9B.
This is no casual acquisition.
Adobe’s going all in on distribution (GEO) and wants control now.

2/ Adobe paid ALL CASH upfront
Everyone sees the shift from Google to Chatgpt,
but paying cash at this valuation signals something different: Adobe believes LLM visibility is the next distribution channel and doesn’t want anyone else to own the “GEO” layer.

3/ Adobe expects to monetize this fast
Cash deals are usually reserved for assets that will drive:
measurable, immediate revenue
direct product integration
a short payback period

Adobe paid a lot for Semrush.
Because in their view, the upside of owning GEO early dwarfs the cost of the acquisition.

My team should feel privileged to work in the office on weekends.


They don’t have to, 
They get to.

As a founder,
I choose the 2 a.m. rebuilds.
The endless emails and fundraising cycles.
The meetings back-to-back.
The weight of decisions no one else can make.

Not because it’s easy.
But because I get to.

“I have to” means pressure.
“I get to” means purpose.

I have to stay up coding another feature.
I get to turn an idea into something real.

I have to pitch again tomorrow.
I get to share a vision that might change how people work.

I have to lead a team through uncertainty.
I get to build with people who believe in the same dream.

That shift doesn’t make the work lighter.
It just reminds you why it’s worth carrying.

how to build $250M lemonade stand in 6 steps:


1. make yummy lemonade (code good product)
2. sell lemonade (sell product)
3. sell 100 lemonades and write down steps, then hire another human to sell lemonade for u (write ur playbook & hire sales)
4. make people know about the lemonade so there is a line out the door (market ur product)
5. line gets long, hire another human to sell ur lemonade (scale sales)
6. line gets short, tell more people about the lemonade (scale marketing)

btw I’m not actually talking about lemonade stands.
im talking about Recall.ai (get pranked).

did i miss any steps?

Our revenue per employee is $1M.
Because I figured out how to scale judgement.

When you’re 2 people, every decision has your fingerprints on it.
When you’re 200, you have to trust people to make calls you’ll never see.

Most founders try to scale process.
But the real unlock is scaling how people think. 

Here’s how we do that at Recall.ai:

1/ Hire for judgement, not IQ
Everyone says “hire smart people.
We look for something narrower:
People who’ve made high-impact calls with 
incomplete information and no safety net. 

You can train skills. 
You can’t train instincts from zero.

-----

2/ Codify your values.
Company values aren’t motivational posters.
They’re a decision-making framework.
When in doubt, they should answer:
Speed or quality?
Autonomy or alignment?
Perfect or shipped?

If values don’t resolve tension, they’re not values.

-----

3/ Run 11-minute daily standups.
Each morning, teams pick one real problem.
I get 60 seconds to say the call I’d make.
The next day, someone else leads.

After a few weeks, you start to see it:
People solving problems the way you would,
even when you’re not in the room.

The goal isn’t just faster decisions.
It’s shared intuition.

-----

You can’t just write judgment into a playbook.
You have to model it until it spreads.

Because your company doesn’t scale with headcount.
It scales with judgment.

I’ve done over 500 founder-led sales calls. These are the 4 things I wish I knew earlier:

1/ Always talk less than the prospect.
Aim for 40% talk time.
Listening forces you to hear the real problem, not just the one you assumed.

2/ Never interrupt.
Not even with “mhm,” “okay,” or “yep.”
Let them get every thought out before you speak.
Show you’re listening with your face, not your mouth.

3/ Qualify hard.
This alone cut our sales cycle in half.
In discovery, we tell prospects every option on the market,
including how they could build in-house.
It builds trust, and it pushes unqualified deals out early.

4/ Stop demo-ing everything.
Our demos are 10 minutes,
Tailored only to their use case.
When you show less, the prospects naturally ask more questions.
Now they’re leading the demo and giving you more signal on what they care about.


This was “Rapid Takes with Recall.ai: Sales Edition”, stay tuned for more 1 minute tips!

Integral

Building next gen Fintech Platform for web3.

MindfulU

MindfulU was on a mission to build a tool to help students improve their EQ. (Funded by LYBI)

Leveraging the latest research in psychology, we developed two models (top-down and bottom-up) to help humans more reliably extract insights from their thoughts and emotions.

Led tech dev -worked with team of PM, designers and researchers to build an mvp solution.
Modelled emotions and thoughts as signals within our graph DB to leverage compression algo to store large datasets (GBs to MBs per user). Also made it easier to surface connections, triggers and patterns as a result (courtesy of some special signal processing + ml algorithms) 
Leverage NLP to construct prompts dynamically as user reveals more data on their emotions.
Built insights module, leveraging three.js and webgl to build delightful dashboards for users to understand and track their emotions and thoughts.

FinSimpl

Lead tech development at Finsimpl.

Designed the first iteration of the site and lead a team to build it out - see www.finsimpl.com
Designed the backend in AWS via SAM & CloudFormation, leveraging lambda's, SQS, DynamoDB, API gateway and AppSync to ensure availability and reliability.
Built the mobile app using react native and released on Apple and Google Play stores.
Setup SOPs, documentation on notion to unblock engineers, with the help of the business team.

Amazon

Worked in personalization, designing and building the next generation of recommenders:
- Researched and scoped a design leveraging a new ML segmenting technology for a hyper-personalized recommender
- Setup routine jobs in Scala to copy over raw data from AWS data lakes into S3 buckets
- Designed and built an AWS step-function workflow to generate structured datasets, set to trigger by the S3 bucket
- Iteratively built algorithms using datasets and round robin processor to generate quality recommendations for customers
- Built and deployed AWS infrastructure using CDK
- Performed E2E and latency testing to ensure quality
- Led a deep dive workshop on a new tool

Wealthsimple

Working for Securities Management to design tooling in Rails and React:
- Designed and built several internal tools used to support trading and investment activities 
- Wrote and helped refactor tests to ensure tools were resilient to errors
- Partook in on-call, aiding in investigating data errors for trade and investing services.
- Led the initial front-end work for the MVP of a new internal dashboard to help on-call engineers

Full Stack Engineer

Initio

Consulted with the Inito Team to help design a local ecommerce app.
Worked with the tech lead to design an MVP using Next and React Native.
Backend was designed to leverage AWS amplify: graphql, S3, Cognito and dynamoDB.

Full-stack Developer

Delivered a new filtering experience on the Buy Again page:
- Optimized system to filter well under 1400ms (faster than filtering on search at the time) by implementing prefetching and client-side filtering
- Leveraged lazy-loading to reduce rendering costs and improve LCP for millions of heavy repeat purchasers
- Launched after successful a/b experimentation and is used by millions of users

Fugro Roadware

Led the development of software to automate data processing:
-Created pipeline to handle post-processing of raw GPS/IMU data and GUI to let the user control pipeline
-Designed a filter to read DB data against raw data and filter out false results
-Developed a QC tool which used a box plot to identify anomalies in the data and overlaid anomalies on Google Earth
-Final Program cut down human involvement from 4 hours to 30 mins
-Developed a download manager for external online files which cut down time from 2 hours to 5 mins

Software Developer

Prometheus Tech inc

Developed a prototype to help reduce carbon emissions from smoke stacks.
Developed a Startup Business Plan to pitch for Investors
Registered with ServiceOntario
Came 4th in PEEL Startup Accelerator Competition

Hi and welcome to my page!

I am a problem solver passionate about designing empathetic, scalable, and reliable software.

Currently focused on building full-stack fintech experiences at companies from pre-series to super-unicorn scale. Exploring AI to build new libraries for a more human-like UIs.

You can find my resume and github linked below!

taplio

The Battle

Amanda Zhu (Hiring SEO Manager)

Krishna Solanki

Followers

Posts

Engagements

Likes

Performance ScoreAmanda Zhu (Hiring SEO Manager)'s LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Amanda Zhu (Hiring SEO Manager)

Krishna Solanki

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);}Amanda Zhu (Hiring SEO Manager)'s LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreAmanda Zhu (Hiring SEO Manager)'s LinkedIn Performance