Dashboard
Last verified 17 Nov · updated daily

Careers at Hiring now

Halluminate

RL environments and benchmarks that teach frontier models to do finance work

7 open rolesSeries A · $30M
Size
5
Backers
OHFYCOC+3
HQ
San Francisco
Industry
B2B

The people who decide who gets hired

  • Hiring managerHiring · 7 open roles

Free account. Look them up, then email them from your own Gmail.

Open roles at Halluminate

7 positions we're tracking, re-checked daily

  • Member of Technical Staff, Research / Post-Training ($200K–$275K, 0.15–0.30%)

    San Francisco, CA, USASeniorFirst seen last month
  • Member of Technical Staff, Platform Engineering ($200K–$275K, 0.15–0.30%)

    San Francisco, CA, USASeniorFirst seen last month
  • Member of Technical Staff, Forward Deployed Engineer ($200K–$275K, 0.15–0.30%)

    San Francisco, CA, USASeniorFirst seen last month
  • Member of Technical Staff, Finance Researcher ($180K–$225K, 0.15–0.30%)

    San Francisco, CA, USASeniorFirst seen last month
  • Data Platform Lead ($200K–$275K)

    San Francisco, CA, USASeniorFirst seen last month
  • Strategic Project Lead ($150K–$225K)

    San Francisco, CA, USASeniorFirst seen last month
  • Chief of Staff ($160K–$210K)

    San Francisco, CA, USASeniorFirst seen last month

Who you'll meet at Halluminate

The Halluminate interview process →

Usually 3 stages over about 7 days. The guide has what each stage tests and how to prepare.

  1. Stage 1Founder or hiring managerIntro Call · 30 min
  2. Stage 2Technical founder or leadTechnical Deep Dive · 60 min
  3. Stage 3Founding teamFinal Round · 45 min

Modelled on how B2B companies with under 50 people interview, not confirmed by Halluminate.

About Halluminate

The product

What Halluminate is building

Halluminate calls itself a data research lab. It builds reinforcement-learning environments and benchmarks so AI models can do knowledge work beyond coding, starting with finance. The product is a simulated workplace: realistic desktop and browser sandboxes where a computer-use agent has to run a real workflow end to end, plus expert-annotated evaluations that grade whether the work is actually correct. The flagship public artifact is the Westworld Finance Diligence Bench. It drops agents into an acquisition due-diligence process built from anonymized private-equity deals: 88 tasks across dynamic desktop environments. When Halluminate ran seven frontier models through it in August 2026, the best average score was 51%. That gap is what they sell. They benchmark a model, find where it breaks on long-horizon financial work, then build training environments aimed at exactly those failures. The buyers are the labs themselves. CEO Jerry Wu says four of the top five closed-source US AI labs are paying customers. His operating principle is a 'Moore's law of environments': environment complexity has to roughly double every 6-8 months or it stops being useful for training. So this is not a dataset business that ships once. It's a research treadmill, and the moat is how fast the team can build harder worlds.

The money

Why this matters

Coding agents got good because code has a cheap, automatic verifier: the tests pass or they don't. Finance has no equivalent. Whether a diligence memo caught the right liability takes a domain expert to judge, and nothing in a frontier model's pretraining corpus looks like the inside of a PE deal room. Halluminate's own benchmark makes the point: the best frontier model scored 51% on its 88-task diligence suite. Labs that want to sell agents into banks and funds need graded, realistic environments, and they can't build them all in house. That's why a nine-person team could reach a mid-eight-figure run rate in roughly a year. Oak HC/FT's Matt Streisfeld put the thesis plainly: once agents get into long-horizon work, 'testing work and specialization will really be key.'

Investors: Oak HC/FT (Matt Streisfeld), Y Combinator, Orange Collective, Heavybit, Angels: individual researchers from Anthropic, OpenAI and Meta

The outlook

Hiring outlook

Explicit and public. The YC Work at a Startup page lists nine open roles, all in San Francisco: four Member of Technical Staff tracks (Forward Deployed, Research/Post-Training, Finance Researcher, Platform Engineering), plus Data Platform Lead, Strategic Project Lead, Chief of Staff, People Platform Lead and a Senior Recruiter. Opening a Senior Recruiter and a People Platform Lead req at nine employees signals an aggressive headcount plan.

Hiring intensity8/8

Halluminate careers: frequently asked questions

How many jobs does Halluminate have open?

7 open roles at Halluminate, last checked November 17, 2026.

Does Halluminate hire remotely?

All current Halluminate roles are based in San Francisco, CA, USA.

What roles is Halluminate hiring for?

Halluminate is hiring across Engineering, Data and Finance. The most recent opening is Member of Technical Staff, Research / Post-Training ($200K–$275K, 0.15–0.30%).

How do I apply for a job at Halluminate?

Press Apply on any role above. You can use Halluminate's own form, or email the person who hires for that team.

Does Halluminate respond to cold emails?

Response rate data for Halluminate not yet confirmed.

Who is the hiring manager at Halluminate?

At this size, hiring is usually run by a founder or department head.

What ATS does Halluminate use?

Halluminate uses Ashby.

How competitive is it to get hired at Halluminate?

<50-person B2B companies see ~50-100 applicants per role in two weeks. The 72-hour window is your advantage.

Where is Halluminate based?

Halluminate is headquartered in San Francisco, CA, USA.

Related

More B2B companies hiring

Other startups we track in the same space.