Hire the engineer who can actually build your retrieval stack.
Pay only when they start.
HatAlign · Relevance is a specialist, done-for-you search practice for AI-native startups hiring in retrieval, RAG, semantic search, embeddings, ranking, recommendation, and agentic search. The roles where every resume lists the same words and only a handful of people can truly do the work. Every candidate is found and screened with AI, then personally assessed and vetted by a former Amazon and Qualtrics applied-science leader, and delivered as a reasoned shortlist. No retainer. A success fee only when you hire someone you love.
No Retainer · No Seats · Pay on Hire.
“ML engineer” hides five different people.
Everyone lists RAG, embeddings, and vector search now; on paper they look identical. A generalist recruiter keyword-matches and can't tell a shipped retrieval system from a weekend notebook. Sourcing tools hand you a longer list, not a better one. And on the model that is your product, a mis-hire costs a quarter of your runway.
Titles mislead
“Applied scientist,” “ML engineer,” “search engineer” mean three different jobs. The right person is defined by what they've owned, not the label.
Resumes are inflated with claims
The signal is in the evidence: shipped systems, papers, code, and whether the claims hold up in a real conversation. Most recruiters can't validate them, and miss matches when the terminology differs.
The best never apply
The engineer you want is heads-down and passive. They ignore agencies and generic InMail. Reaching them takes a credible, specific reason.
You keep the pitch and the final call. I carry the rest.
A recruiting subscription still leaves the sourcing, screening, outreach, and chasing on your plate. This doesn't. Here is the split, end to end.
- 1Define & scope the roleTOGETHERA 30-minute intake; I decompose the vague title into the real capabilities.
- 2Source candidatesHATALIGNAI-native search across the open web, papers, and code, including people who never applied.
- 3Screen & rankHATALIGNDeep, domain-aware screening, then my own expert assessment of every candidate.
- 4Vet the shortlistHATALIGNA real conversation with each finalist and a light technical check before you see anyone.
- 5Outreach & warm introsHATALIGNI reach out with a specific, credible reason and hand you a warm introduction.
- 6Interviews & referencesYOUYou run them. I supply the exact questions and the risks to probe.
- 7Pitch & closeYOUYou sell the vision best. I support with what moves this specific candidate.
Most of the work comes off your plate. What stays is selling the vision and making the final call.
Founders and engineering leaders
at AI-native startups
Seed to Series C, making a critical hire where getting it wrong is existential:
AI for reach and speed.
A domain expert for the judgment.
Scope the role
A 30-minute intake to decompose what you actually need behind the title. We align on 2-3 example profiles before any full search.
Find & screen
AI-native search and deep screening across the web, papers, and code. Every candidate is assessed for technical readiness, experience alignment, professional maturity, AI readiness, and early signs of culture fit before anything reaches you. See what we evaluate →
Vet
A real conversation with each finalist plus a light technical check, informed by staying current as an active reviewer in the field I recruit for.
Decision-ready shortlist
A small, ranked list where every candidate arrives as a reasoned case: why they fit, where the risks are, and what to verify.
You decide
Interview the right few with warm intros. Pay a success fee only when you hire someone who starts, backed by a guarantee.
Not a cheaper agency.
A different kind of partner.
| Generalist recruiter agency or in-house | DIY founder + AI tools | HatAlign · Relevance | |
|---|---|---|---|
| Can they evaluate retrieval / ML depth? | No, a generalist.You end up re-screening every candidate yourself. | Only if you do it yourself.On top of everything else. | Yes.Candidates arrive pre-evaluated by someone who's built these systems. |
| Is each candidate personally vetted? | Rarely, and never technically.You discover gaps in the first interview. | No.You're the only filter. | Yes.Every finalist gets a conversation and a technical check before you see them. |
| Do you see the reasoning? | No, a stack of resumes.You're guessing which to read first. | No.A longer list, still no reasoning. | Yes.Each candidate comes with why they fit, strengths, risks, and what to probe. |
| Your cost and risk | 20-30% per hire (agency), or a full salary at volume (in-house).You pay whether it works out or not. | Your hours + tool fees.Regardless of outcome. | |
| Specialization | Everything.The recruiter can't tell RAG from a chatbot wrapper. | Whatever you search.No domain filter. | Only retrieval / RAG / ranking / multimodal / agentic.That's where the depth is. |
Already have a recruiter? An in-house or contract recruiter is still a generalist who can't evaluate a retrieval or ranking engineer, and a full recruiter (~$95-165K/yr, loaded) only pays off once you're making four or five hires a year. For one or two critical ML hires, a specialist on a success fee is both cheaper and deeper.
The honest exception: the one thing a specialist agency still offers is a human closer for the very hardest senior hire. For most roles you won't need it. The strongest passive candidates respond to a founder and a technical peer, not an intermediary. Because I'm an applied-science leader myself, I can step in as that peer to help close when a search is the exception. I'll tell you honestly when it is.
I've built these systems,
and hired the people who build them.
HatAlign · Relevance is led by Prashant Shiralkar. I've hired and evaluated this talent at Amazon and Qualtrics, built the same systems myself, and stay current as a reviewer for top AI venues. I evaluate candidates as a technical peer and a hiring leader, not a keyword matcher.
No retainer. No upfront fee. Pay only when you hire.
One specialist role, worked end to end: found, vetted, and delivered as a reasoned shortlist. You pay a 20-25% 15% introductory success fee on first-year salary, only when you hire someone you love, backed by a replacement guarantee. Lower because the sourcing is automated to surface the right few, not the longest list. The judgment is mine. Early clients lock in this rate.
Straight answers
How is this different from a sourcing tool?
A sourcing tool hands you a longer list and leaves the screening to you. This is done for you, end to end. HatAlign's AI handles the reach and first-pass screening. I then add my own expert assessment and a real conversation with each finalist. Nothing reaches you until it's a decision-ready shortlist, not a haystack.
What does the vetting involve?
I speak with each finalist and run a light technical check on the specifics of the role. Because I actively review submissions for venues like ACL, COLM, EMNLP, and VLDB, and build these systems myself, I stay current on what genuinely strong work looks like right now.
What exactly do you evaluate?
Every candidate is assessed across five areas before the shortlist reaches you. This is how we identify strengths, gaps, and specific areas for you to probe during interviews.
Functional and technical fit. Do their hard skills match? Can they handle the complexity and scale this role demands?
Track record and trajectory. Have they delivered measurable impact? Is their career progression logical and stable?
Professional maturity. Can they own the scope, lead at the expected level, and communicate with stakeholders?
Evidence quality. Are the claims specific and verifiable? Shipped systems, papers, quantified results, not buzzwords.
AI resilience. Which parts of their experience are durable and matter for this role, and which are being commoditized? Scored against human-rated work tasks and real-time market signals.
If your role has requirements not covered above, we accommodate those too. The evaluation is tailored to what you actually need, not a fixed checklist, but these five areas are what we have consistently seen to matter most in the modern hiring process for a reliable hire.
What does it cost?
20-25% A 15% introductory success fee on the hire's first-year base salary, payable only when they're hired and start. No retainer, no upfront cost. Early clients lock in this rate.
What's the guarantee?
If a placed candidate doesn't work out within an agreed window, the search is redone at no additional fee. Terms are set per role in the engagement agreement.
Who makes the pitch to candidates?
I find, assess, and vet each candidate, then make a lightweight pitch on your behalf to get them excited and engaged about the role. You sell the vision directly with a warm introduction and a specific, credible reason to talk. Founders close the best passive candidates, not intermediaries.
Do you recruit for other roles?
Adjacent AI/ML roles, sometimes, by referral. But retrieval, RAG, ranking, multimodal, recommendation, and agentic systems are the deliberate focus. That's where the depth and the network are.
Tell me the role. I'll tell you who's actually out there.
A 30-minute call to scope one hard hire. No retainer, no obligation. You only pay if you hire someone you love.