Careers at AfterQuery: Teams, Pay and How to Get Hired

Who Gets Hired at AfterQuery

A roughly 30-person team hit $100 million in annualized revenue by April 2026. NVIDIA had already named AfterQuery the sole data vendor in a technical report. Motif Technologies, a Korean AI lab, was a customer. The product worked.

AfterQuery runs a core team of roughly thirty in San Francisco, including engineers, researchers, and strategic-projects leads, alongside an expert network of nearly 100,000 verified professionals paid by the hour to produce reasoning traces that become training data. The two tracks operate on different rails: the corporate side is onsite, full-time, and pays $200k–$450k base with equity; the expert side is remote, contractor-only, and pays $40–$250 an hour, according to AfterQuery's experts portal. Candidates who conflate them waste interviews preparing for the wrong process.

The core team orbits three co-founders who met before founding. Mateega (CEO) worked at Silver Lake, Morgan Stanley, Meta and Google. Georgescu came from Citadel Securities, then Meta and Google. Tang sold a startup with Mateega in high school and roomed with him at Wharton. All three entered Y Combinator's Winter 2025 batch. YC partner Gustaf Alströmer said in September 2026 that AfterQuery was the fastest launch-to-unicorn in the accelerator's history. Zero G Talent's board shows 11 salaried postings with a median band of $200k and a ceiling of $400k. Roles cluster in four functions:

Engineering splits between platform infrastructure (systems that ingest, version and serve expert trajectories) and applied AI, where researchers turn trajectories into training runs and evaluation rubrics. AfterQuery's blog reported Terminal-Bench 2.0 showed a 5x score improvement; its data shows a GDPval ablation added 11 points on Nemotron 3 Ultra. Strategic Projects runs like a chief-of-staff layer: leads own engagements with frontier labs (the NVIDIA partnership ran through here), associates handle operations. Marketing is a single-lead role; revenue comes inbound.

The expert network onboards verified professionals as independent contractors — a 3D Designer/Blender Expert at $40–50/hr, an Actuary Expert at $80–150/hr, an Audio Recording Contributor for Australian English at $50/hr. They don't attend standups. They don't get equity. They supply the "tacit knowledge" the thesis says lives "in the space between keystrokes, in the pause before a diagnosis." The expert portal handles intake.

No published checklist exists for core-team qualifications, but the roles and founder backgrounds set the bar. Engineering candidates need production-grade systems experience at scale. Research hires need fluency with RLHF, chain-of-thought supervision and agent environments. Strategic Projects leads translate lab requirements into data specs and manage expert-network output quality. Associates earn the same $200k base as leads and operate with minimal supervision. Headcount is a constraint, not a target. The team stays small while the expert network scales. Every core hire must replace a process, not just execute one.

What AfterQuery Pays

AfterQuery's compensation leads the early-stage market, reflecting a $3.2 billion valuation. Bands often span $100k or more, and variable cash can multiply earnings several times for top performers. Contractor roles run on different economics. Benefits look generous in summaries but stay opaque in fine print.

Pay by function

AshbyHQ postings and Zero G Talent's board data track live listings. Engineering and technical-lead roles cluster at $200k–$450k base with equity and bonus on top. Go-to-market and operations roles start at $150k.

AshbyHQ postings say total cash compensation regularly exceeds base by 2–3x, and 4–5x for top performers — variable pay is structural.

Equity: what's missing

Every full-time posting mentions equity but none disclose grant sizes, vesting schedules or strike prices. The careers page promises "meaningful equity" and the chance to "shape the engineering organization and lead major technical initiatives" as the company scales. Backed by Altos Ventures, BoxGroup and angels from DeepMind, OpenAI, Anthropic, Meta Superintelligence Labs and Microsoft AI, grants likely match Series B–C benchmarks. Candidates should request the current 409A valuation, outstanding share count and their grant's percentage of fully diluted equity before signing. BuiltIn notes that benefit details like premiums, match %, and eligibility can shift; equity terms may be similarly fluid.

Benefits: generous summaries, thin details

For full-time employees, AfterQuery advertises a package heavy on lifestyle perks:

• Daily meal stipend via UberEats
• Commute stipend via Uber credits
• Monthly wellness stipend covering Equinox membership
• Medical, vision, and dental insurance (described as "comprehensive" but without premium splits, network details, or HSA/FSA confirmation)
• 401(k) match (percentage and vesting undisclosed)

BuiltIn's August 2026 FAQ says employee materials indicate medical, vision and dental insurance are provided but notes "scant, independently verified plan specifics." It advises candidates to "secure written benefit summaries and amounts before committing." Contractor and expert roles lack healthcare or retirement coverage; their materials focus on rates and payouts, not employer-sponsored benefits.

Contractor pay: different rails

Expert-track contractors earn $40–$250/hr per the experts portal; research interns and fellows on AshbyHQ show $8k–$16k/month. Per-task payments of $1k–$10k appear for scholar roles. All are approval-based. BuiltIn warns that "pay is often contingent on lengthy or unclear acceptance criteria, with work sometimes archived or left on hold without payment. Task caps and approval-based compensation can leave actual earnings below headline rates." Slow review queues and spotty project availability create unpredictable cash flow — unlike salaried roles where bonuses tie to defined cycles.

Before you sign, ask

1. Written benefits summary: premiums, 401(k) match %, vesting, eligibility waiting periods.
2. Equity grant details: shares, percentage, current 409A, exercise window on departure.
3. Variable pay mechanics: bonus formula, performance metrics, payout timing, clawback provisions.
4. Role classification: full-time employee vs. expert/contractor determines the entire benefits stack.

AfterQuery pays like a late-stage company while operating with early-stage opacity. The headline numbers are real and competitive; the fine print is where the negotiation lives.

How AfterQuery hires

AfterQuery uses Ashby for recruiting. Applications land in arrival order; recruiters review sequentially, so early applicants get more attention. Scoutify's board shows 14 open roles as of late May 2026, all in San Francisco — recent postings include Talent Coordinator, Technical Recruiter, Business Talent Acquisition Lead, People Programs Lead and Customer Success Lead. No H-1B petition history exists; international candidates needing sponsorship should clarify eligibility first.

For software engineering roles, the loop opens with a recruiter screen and one to two technical rounds on data structures and algorithms. InterviewSense rates difficulty medium (5/10 across six reports). Dataford's breakdown shows AI evaluation systems, data pipelines, distributed systems, reinforcement learning and RLHF all above 85% coverage; RLVR sits at 78%. Expect questions mirroring the actual workload: building evaluation harnesses, scaling data pipelines, hardening infrastructure for frontier model labs.

Expert-track roles (finance, law, medicine, engineering) use an AI-moderated interview. A 2025 YouTube walkthrough shows a 19-minute session with three question sets; candidates can retake up to three times. The system requires continuous camera facing (pausing triggers auto-submit) and thoughtful answers; "very short interviews may not be considered complete." The platform says responses are only for candidacy assessment, never for model training. Applications require a LinkedIn profile, proof of PhD or equivalent, and links to publications with the candidate's name. One efficiency: a finance expert interview recording can be reused for other finance postings.

Recruiters screen for three things: verified domain expertise (credentials, publications or comparable track record), communication clarity under time pressure, and alignment with a high-ownership, direct-feedback culture. The company calls itself a fast-paced applied research lab for foundation model developers; Scoutify's tags highlight "high standards" and "direct feedback." Glassdoor's six reviews are mostly negative (67%). One engineer cited "excellent opportunities for engaging in open source projects"; another flagged "project reviews can be time-consuming, which may impact workflow."

A strong application leads with concrete proof: a GitHub repo showing production-grade infrastructure, a published paper, or a portfolio of expert-level deliverables. Timing matters — applying within hours puts you at the top of the Ashby queue. Disqualifiers: visa dependency (no sponsorship track record), inability to sustain camera engagement, vague responses suggesting domain unfamiliarity, and evidence of outsourcing the interview (the platform records audio and video). Mateega said revenue reached several times $100M ARR by July. Hiring velocity is high; the bar stays narrow. They need people who ship evaluation systems and data infrastructure that move benchmark scores for frontier clients like Nvidia, Thinking Machines Lab and Legora.

Office: San Francisco only

AfterQuery is headquartered in San Francisco, where it was founded in 2025 and completed YC W25. The aVenture profile lists the HQ as "San Francisco, CA, US" — no neighborhood, floor or building name. Headcount ranges from 102 (PitchBook) to 167 (aVenture). The company is past the sublet phase but not at campus scale.

Every Zero G Talent listing is tagged San Francisco. Six live listings (Strategic Projects Lead, Technical Strategic Projects Lead, Senior Software Engineer (Infrastructure & Platform), Marketing Lead, Software Engineer (Applied AI/Platform) and Strategic Projects Associate) all specify San Francisco. None mention remote or hybrid options. Salary bands match San Francisco rates. The pattern points to an in-office default for active roles.

No public hybrid or remote policy exists. Generic SHRM and BuildRemote templates appear in search results — not evidence of AfterQuery's policy. Without a first-party statement, job postings are the strongest signal: if hybrid or remote were standard, a listing would flag it. Silence points to an office-centric model. Expert-curated datasets and RL environments for frontier models need close hardware collaboration, secure pipelines and low-latency loops hard to replicate over Zoom.

San Francisco's commercial market remains a tenant's market. PropertyShark lists over 2,000 office spaces for lease. A Series A company of this size can negotiate favorable terms, often in older buildings with fewer amenities than pre-2020 trophy towers. Expect a functional, engineering-oriented workspace, not a showcase office.

Applicants should plan to work from San Francisco. Hiring, compensation and team size orbit the HQ. If you're not in the Bay Area or willing to relocate, current openings offer no remote path. That may change after Series B, as many YC companies do, but for now the office is the only option.

The profile that lasts

People who last share a profile: they blend elite technical execution with professional-domain expertise and treat ambiguity as a design constraint. The core team has grown to about 30, spanning engineering, frontier data research, post-training, strategic projects, growth and sales, and runs almost entirely onsite in San Francisco. That density is intentional. Founders Mateega, Georgescu and Tang built careers at Meta, Google, Citadel Securities, Silver Lake and Morgan Stanley before launching in January 2025. Their first product (AI agents for financial analysis) failed because models lacked professional work data. They pivoted to supplying it. That sequence (rapid commitment, fast failure, hard pivot to infrastructure) recurs in people who succeed. The team includes former Citadel quants, Meta and Google infrastructure engineers, and operators who have shipped at scale. A Glassdoor review called the process "inexperienced and not especially transparent" — a filter, not dysfunction. The company doesn't over-explain. It expects candidates to reverse-engineer the problem space from public signals — Nvidia technical reports, Legora case studies, the GDPval benchmark (over 1,300 real tasks across 44 occupations and 9 industries), the BAR legal-reasoning benchmark with Legora (5,161 cases across 28 practice areas), and arrive with a point of view.

Two distinct paths exist; conflating them wastes time. The Experts contributor program (experts.afterquery.com) pays $40–$50/hr for general work and up to $200–$250/hr for specialists, per AfterQuery's experts portal; it's remote, flexible, and open to professionals with 1–3 years' experience (no PhD required). The corporate track (14–38 open roles) is onsite, full-time, and demands the velocity of a profitable Series A lab. People who thrive on the corporate side treat the expert network as a product surface to instrument, not a labor pool to manage. They build evaluation rubrics, agent environments (API/MCP), computer-use trajectory capture, and RL reward models that make data useful to Nvidia or Legora.

Reddit and Glassdoor threads flag friction: payment delays for contributors, opacity in recruiter communication, skepticism that "software engineer" roles are data-annotation jobs in disguise. The pattern suggests a company outpacing its candidate experience. People who need structured onboarding, clear career ladders or remote flexibility will churn. People who want to own a slice of the data-to-model pipeline (from expert recruitment software to post-training validation) and can operate without a spec will compound. The bar isn't credentials; it's whether you've built something that required stitching together domain experts, ML infrastructure and measurable model improvement.

Working in AI? Zero G Talent tracks the openings: see every open AfterQuery role, browse AI jobs, the companies hiring, and the people building the field.