AI-native search Venture-backed startups SF · NYC · Toronto

The engineers building the frontier aren't applying. We find them anyway.

StackedSP is the search partner for venture-backed companies building with AI. Forward-deployed engineers, AI engineers, software engineers and the leaders who run them. The whole market mapped by our own system, every shortlist judged by people who have placed engineers for more than a decade. Our system finds them. Our people win them over.

stackedai os · searchlive
>
sources public + privateloop outcomes feed back
Our partners

We take on partners, not clients, and we stay through the rounds that follow. Seed to Series D to category-defining company, with the funds that back them. Increasingly that includes the established companies our startup partners sell to, now hiring the way startups do.

Teams we've built
Their investors
What we place

The roles that take AI from demo to production.

01 / forward-deployedHardest to find

Forward-deployed engineers

One of the rarest profiles in AI right now: engineers who can sit with a customer, tell which request is the product and which is a distraction, and ship the integration the same week. We keep a specific map of who has done this and who only says they have.

FDESolutions engineerImplementation leadCustomer-facing AI
02 / ai & ml

AI and ML engineers

Applied AI, research engineering, machine learning platform and inference infra. People who have shipped models to production and can write the eval that proves it. We screen for taste, which is harder to find than skill and shows up in what someone chose not to build.

Applied AIResearch engML platformLLMsEvalsTasteInference
03 / software

Software engineers

Founding engineer through staff. Full-stack, backend, infrastructure, product-minded builders who want to own the whole problem and ship weekly.

FoundingBackendFull-stackInfraStaff
04 / leadershipExecutive search

Engineering leadership

CTO, VP Engineering, Head of Engineering, Head of AI. Executive search run by people who can tell the difference between a leader who scaled a team and one who was present while it scaled.

CTOVP EngineeringHead of AIHead of DataChief Data & AI Officer

When the team around the engineers needs to keep up, we place those roles too.

GTM engineersAI product managerProduct designData scienceDevRelSales engineeringGTM leadership

AI can generate a hundred ideas an hour. The hard part is knowing which three are good.

The engineer

The hire every AI company wants right now is an engineer with taste. Someone who can say what good looks like before the code exists, and write the eval that proves it.

The search

Hiring works the same way. Our system generates the candidates. Our people decide which ones are good. Every shortlist is an eval we are willing to be judged on.

Who we work with

Two starting points. One bar.

01Building AI-native

Venture-backed companies born on this stack, seed to Series D. Our sweet spot.

02Becoming AI-native

Established companies that have decided to make the shift and need a Head of AI, a Head of Data, and engineers who have shipped this before. Real mandate, real budget, a blank page.

Different cultures, different starting points. The bar for who gets hired is the same.

StackedAI OS

The operating system behind every search.

The learnings from every search we have run, the public record of who is building what, and your search requirements, in one system. Agents do the reading and the ranking. Recruiters do the judging and the closing. More than a decade of outcomes are already in it, so your first search starts with everything we have learned.

fig. 01 · how data moves through a searchrunning
StackedAI OS: public signals, first-party data and your search requirements flow into one talent graph; the champion score, matching engine and critic agents rank the market; a recruiter judges the shortlist; outcomes feed back in OUTCOMES FEED BACK IN SOURCES ONE DATA LAYER INTELLIGENCE JUDGMENT Public signals profiles · code · paperstalks · launches First-party data a decade of searches,interviews, outcomes Your search requirements the bar · the stage · the comp Talent graph every candidate, conversationand outcome in one place enrichment attributes history Champion score four pillars,refined over a decade Matching engine ranks the marketagainst your bar Critic agents review every messagebefore it goes out Recruiter judges ·interviews ·closes
Scroll sideways
01 · sources

Public signals, first-party data, your requirements

Profiles, code, papers, talks and launches from the public web. More than a decade of our own searches, interviews and outcomes. And the bar, the stage and the comp you set with us.

02 · one data layer

The talent graph

Every candidate, conversation and outcome in one place, enriched from public and private sources and broken into the attributes that predict success in the role. Nothing important lives in someone's inbox.

03 · intelligence

Champion score and matching

Our Champion methodology scores every engineer on four pillars refined over more than a decade, and looks for taste: what they shipped, how they measured it, what they chose not to build. The matching engine ranks the whole market against your bar and shows its reasoning.

04 · judgment

The recruiter

A recruiter who knows the market makes the call, runs the interviews and closes. Every outcome goes back into the graph, and your search runs on everything it has already learned.

Faster first shortlistWider talent poolFewer wasted interviewsOutreach timed to signalsA better candidate experienceMore time with the best peopleHumans and agents work as one team
That is the science

This is the art.

Finding champion talent is a science. Persuading them to make a change is our art.

RelationshipsThe best people pick up when we call

More than a decade of relationships with engineers and leaders at frontier labs like Anthropic and OpenAI and at the startups moving fastest. Built one conversation at a time, long before there was a role to fill.

JudgmentKnowing who is really movable, and why

A scorecard tells you who is strong. A person who has run hundreds of these searches tells you who will actually say yes, what it will take, and when to walk away.

ClosingWhere searches are won and lost

Offers, counters, the partner's cold feet and the candidate's. We stay in the room until the signature, and after.

The roadmap does not wait

Every week the seat stays empty, the roadmap slips.

The system is already trained on more than a decade of searches. Yours starts at full speed on day one.

How a search runs

Built to move at the speed of your roadmap.

Four stages. The system does the science: mapping the market and scoring every candidate. Our people do the art: calibrating the brief, persuading the best to listen, and closing.

01science + art

Calibrate

A working session on the role, the stage, the bar and the market. We tell you whether the role as written is findable, what it pays, and what to change if it isn't.

More than a decade of search data
02science

Map

The system maps the whole market for the profile: who is building what, where, and who is realistically movable. You see the map, not only the people on it.

talent graph + enrichment
03science + art

Assess

Every candidate is read and scored before a human spends a minute on them. The humans then do the part that matters: structured interviews against your bar, references, judgment.

champion score + matching
04art

Close

We run the offer with you, handle the counter and stay close after the start date, so the hire sticks. Most clients come back for the next one.

learning loop
The loop · recalibrated every week

A search is never a straight line.

The first shortlist is a hypothesis. Your feedback on it tightens the bar, redraws the map and sharpens the message, and the search iterates week over week until it closes. Calibrate, map and assess run as a loop, not a sequence.

Proof

Founders who bet on the people we found.

StackedSP has delivered world-class talent across multiple functions and geographies, with outsized impact.
Nicolas DesmaraisChairman & CEO, AppDirect
They built our teams in engineering, product, design, marketing and operations, and helped us launch into new markets.
Alyssa RavasioFounder & CEO, Hipcamp
StackedSP consistently delivers engineering talent at a high bar of excellence.
Vivek SubramanianFormer SVP Engineering & Product, Sonder
About

Operators first, recruiters second.

StackedSP was founded by Ilan Saks and Jared Saks and works out of Toronto with venture-backed companies across North America, mostly in San Francisco, New York and Toronto. Ilan started his career at a seed-stage venture fund and still invests in the companies and funds the firm works alongside.

The team is eight people, on purpose. You work with the people who run your search from the first call to the signed offer, and every one of them lives inside the same system you saw above.

We also run Unstacked Startups, a podcast and newsletter for founders and operators scaling venture-backed companies. It is how we stay close to the people we place, and how we give back to the community that hires them.

San FranciscoNew YorkToronto
Our values
  1. Go above and beyond

    We exceed expectations by always taking the extra step.

  2. Be fierce

    We face the impossible with energy, grit and cleverness.

  3. Take ownership

    We own our actions and don't make excuses in the face of great challenges.

  4. Be humble

    We collaborate to achieve the mission. We listen and never stop learning. We don't take credit, we give it.

  5. Stay positive

    We strive to affect change through positive thoughts and actions. We present solutions, not problems.

Start a search

Let's build together.

Category-defining companies are built one hire at a time. We would like to build yours with you.

A thirty-minute call: whether the role is findable, what the market pays, and how long it will take.

Book a call