Skip to main content

Engineering · Senior

AI Research Engineer

AI Research owns the recognizer training pipeline, the consensus reconciler tuning, and the active-learning loop. Heavy emphasis on production correctness over leaderboard accuracy — every model that ships has to survive a senior estimator's challenge.

Location

San Francisco, CA · or remote

remote

Team size

ML research + backend team; specific headcount discussed in interview

Department

Engineering

What you'll do

Day-to-day responsibilities

  • Improve symbol recognizer accuracy across trades
  • Tune consensus reconciler signals (vector grounding, library boost, trade penalty)
  • Design + run active-learning experiments
  • Build ML evaluation infrastructure
  • Publish internal tech reports + occasional external blog posts

What we're looking for

Required experience

  • ML engineering background (CV preferred — object detection, segmentation, OCR)
  • Experience shipping ML to production, not just notebooks
  • Strong Python + at least one of PyTorch / TensorFlow
  • Comfort with classical CV alongside DL — the right tool wins
  • Interest in domain-specific applications (this is not a leaderboard chase)

Nice to have

Bonus experience

  • Construction tech, AEC, or document-AI background
  • Active-learning + uncertainty estimation expertise
  • PhD or equivalent industry experience

Compensation + benefits

What you'll get

  • $220K-$280K base + equity
  • Full benefits stack
  • Conference + paper-publication encouraged
  • GPU compute budget

Hiring process

Five-step process, 4-6 weeks end-to-end

  1. 01

    Application

    Send a resume + a one-paragraph note on why this role. We read every email — no ATS black hole.

  2. 02

    Recruiter screen

    30-minute call to align on logistics + level + comp.

  3. 03

    Hiring manager interview

    60-90 minutes — deep on background, what you'd want to own first.

  4. 04

    Technical / role-specific loop

    2-3 sessions covering the hardest parts of the work. We don't trick — every question maps to something you'd actually do.

  5. 05

    Offer + reference checks

    We move fast on offers. References are quick — we ask three questions, not thirty.

Apply

Send us a note about AI Research Engineer

Email careers@omnitakeoff.com with a resume and a one-paragraph note on why this role. We respond within 5 business days.

AI Research Engineer — OmniTakeoff Careers