Forward-Deployed Engineers · Agentic Systems Builders · AI Platform Leaders · Founding Engineers

The search and verification partner for FDE and agentic systems hiring.

Every engineer’s résumé says agentic systems now. We can tell who can actually build, deploy, and earn trust with customers — from 500+ engineering hires through NVIDIA’s AI infrastructure shift.

“Jason’s approach to talent recruiting is state-of-the-art and game-changing.”
— Jensen Huang, Founder & CEO, NVIDIA

That search judgment was built through 500+ engineering hires at NVIDIA during the AI infrastructure shift — Senior, Senior Principal, Staff, Distinguished, and Senior Distinguished Engineer roles across AI, GPU cloud, storage, and platform.

The signal problem

The titles are messy. The real difference isn’t on the résumé.

  • Forward-Deployed AI Engineer
  • Agentic Systems Engineer
  • Applied AI Engineer
  • AI Platform Lead
  • Founding Engineer
  • AI Engineer
  • Distinguished Engineer
  • ML Platform
  • LLMOps

A strong résumé and a coding screen still miss the real difference: who has actually shipped agentic systems against real users, real latency, real cost — and who is rehearsing the vocabulary. That gap is what we close before you sign an offer.

Deployment fit

Technical proof is not enough.

FDEs do not enter clean systems. They enter real teams, internal politics, customer pressure, technical debt, and environments where people may feel threatened by the mandate.

The hire has to be more than technically strong. They need the judgment, temperament, and trust-building style to succeed inside the system they are being deployed into.

RunRetained assesses deployment fit before the offer: whether the candidate’s operating style matches the team, customer environment, pace, ambiguity, and mandate.

Your team invests under 30 minutes. You get a decision-grade read before the offer, not after the burn.

Project Atlas

Every engagement produces the same six artifacts.

You see the work, not just the verdict. Below: the working files from one VP, AI Infrastructure search. Identifying details anonymized.

PROJECT ATLAS·MANDATE BRIEFSearch ID: ATLAS-VP-AI-INFRA-2026-Q2 · Opened: April 14, 2026 · Partner: J. La Barbera

Role

VP, AI Infrastructure

Reports to: CTO · Direct reports: 4 (growing to 9 within 12 months)

Scope: Full ownership of training, inference, and developer platform

Calibration parameters

Must have

  • Production experience scaling >1,000 GPU clusters
  • Built or run a multi-tenant inference platform at >$10M ARR
  • Hired and held 4+ senior IC/EM staff through hypergrowth

Strong have

  • NVIDIA, hyperscaler, or AI-native infra background
  • Public technical voice (talks, papers, GitHub)

Dealbreakers

  • Pure research background with no production scar tissue
  • Anyone who has not survived an on-call rotation in the last 3 years

Compensation envelope

Base$325K – $375KEquity0.4% – 0.7% (Series B refresh post-close)Sign-onNegotiable up to $250KY1 target$1.1M – $1.4M

Process commitments

Slate target5 candidatesSlateMid-search slate checkpointFirst offerOffer-stage milestoneCloseClose-stage milestone
Confidential · RunRetained · Project Atlas · 14 Apr 2026
PROJECT ATLAS·MARKET MAPSnapshot: early in search

Segment 1 — AI-native infra leaders

Identified47Qualified22Contacted18In conversation9

Cohort leads: A1, A2, A3, A4, A5, A6

Strongest pool. Two Staff-level candidates already pre-IPO.

Segment 2 — Hyperscaler platform directors

Identified31Qualified14Contacted11In conversation4

Cohort leads: B1, B2, B3, B4

Comp delta is the friction. RSU-heavy packages competing.

Segment 3 — PE-backed infra CTOs

Identified18Qualified6Contacted5In conversation1

Wrong altitude. De-prioritized as primary segment early in mapping.

Confidential · RunRetained · Project Atlas · 23 Apr 2026
PROJECT ATLAS·CANDIDATE DEFENSECandidate A2 · Status: Strong Yes · Recommended for slate

Candidate: A2 · Current: Staff Engineer, AI-native unicorn
Status: Strong Yes · Recommended for slate

Why them — evidence

  • Built and ran the training cluster from 800 → 6,400 GPUs over 22 months
  • Authored the company's internal inference-cost reduction program (43% reduction, two quarters)
  • Hired 6 of 8 senior ICs currently running the platform
  • Operating altitude: presenting at board level despite IC title
  • Public technical voice: 2 conference talks, active on hardware forums

Risk surface

  • First-VP move. Mitigation: already operating at scope; CEO reference confirmed
  • Comp gap: $4.2M unvested. Mitigation: sign-on + accelerated refresh modeled
  • SF Bay-based, role hybrid. Mitigation: confirmed willingness to travel monthly

Reference signal — back-channel, 4 sources

Direct manager: "If she leaves, we're in trouble. She's the operating spine of the platform team."

Peer Staff Engineer: "The most clear-headed infra leader I've worked with in ten years."

Future-proof lens

24–36 month scale view: operating altitude already at VP. Pattern of compounding scope ownership. Risk of plateau low; risk of outgrowing the role exists if Series C scope does not expand.

Recommendation

Strong Yes. Lead candidate. Move to founder interview at next-stage milestone.

Confidential · RunRetained · Project Atlas · 06 May 2026
PROJECT ATLAS·SLATE OF FIVEMid-search slate checkpoint · Slate delivered to client
A2Staff Eng, AI-native unicorn · 800→6400 GPU build · First VPStrong Yes
A4Sr Director, hyperscaler · Multi-tenant infra at scale · Comp gapYes
A1Staff Eng, AI-native unicorn · Inference cost program lead · Comms styleYes
B1Director, hyperscaler · 12K GPU production fleet · Risk-averseQualified
A6Principal, AI-native unicorn · Built platform from zero · Earliness in roleQualified
Confidential · RunRetained · Project Atlas · 19 May 2026
PROJECT ATLAS·WEEK 5 MEMOSent: Monday, May 19, 2026 · 8:00am PT

Velocity

Outreach sent42 ▲ 6First conversations11 ▲ 2Deep-dives5 ▲ 1Defense briefs3 ▲ 1Slate-track7

Conversions

Outreach → conversation26% (target 22%)Conversation → deep-dive45% (target 40%)Deep-dive → defense60% (target 55%)

Slate health

On track against agreed search cadence. Lead: A2 (Strong Yes, founder interview Wed). Backup depth: A1, A4 advancing. B1 and A6 in evaluation.

Decisions needed

  • Equity refresh structure for A2 — CFO sign-off by Friday
  • Reference release for A4 — pending current employer notification window
  • Confirm: open to remote-first for the right candidate?

Next week commitments

  • Founder interviews: A2 (Wed), A1 (Thu)
  • 8 net-new outreach in Segment 1
  • Defense briefs: A4, B1
Confidential · RunRetained · Project Atlas · 19 May 2026

Same artifact discipline — mandate brief, market map, candidate defense, slate, weekly memo — across three engagement shapes.

Founder

Built by the person who delivered 500+ engineering hires at NVIDIA.

Jason La Barbera built RunRetained from 25 years finding builders and leaders for technical companies — including six years at NVIDIA during the AI infrastructure shift, where he delivered 500+ engineering hires across Senior, Senior Principal, Staff, Distinguished, and Senior Distinguished Engineer roles, plus 100+ executive hires across AI, GPU cloud, storage, and platform. He built the firm to be the one he wished existed when he was the buyer.

Confidentiality is the default. Named references available on request, post-NDA.

  • One partner. A visible operating system. Judgment stays with Jason.
  • Compounding. Your second search starts with the calibration earned on the first.

If an FDE, agentic systems builder, AI Platform Lead, Distinguished Engineer, or founding engineer is on your slate this quarter, start with one conversation.

Jason replies within 24 hours.

Confidentiality is the default. Sensitive engineering hiring questions, replacements, and finalist assessments are handled directly by Jason and shared only with agreed stakeholders.

FDE and agentic systems search. Hire. Verify. Map.Confidentiality is the default.jason@runretained.com© RunRetained 2026