Job description

ABOUT TETSUWAN SCIENTIFIC Few fields touch human life as directly as biology does. But the pace of biological research is limited by how slowly and inconsistently experiments can be run by hand. We're building the infrastructure to make experimentation faster, more reproducible, and available to any scientist with questions worth asking. For decades we've had incredible robotic tools capable of executing experiments without countless tiring hours of pipetting by hand and these instruments are more precise and accurate than manual methods. But the engineering and validation time required to automate them means they're rarely used outside of the most repetitive, high-throughput research tasks. Our team has built a user-friendly editor and compiler that ingests a scientist's protocol description, converts it into a structured high-level format, and then automatically compiles it into executable code for these robots. See a demo here: tinyurl.com We're a small, quickly growing team based in San Francisco. Joining now means shaping the product, the team, and the company at a stage where your decisions matter. THE ROLE You'll be joining a team of three software engineers, an automation engineer, and working directly with the founders and our designer. The monorepo codebase spans a compiler (OCaml), backend services (Python), and a protocol editor (TypeScript/React) and we expect you to grow into working across the full stack. Our engineering challenges go well beyond a typical web app or agent system, involving compiler design and the messy quirks of physical lab hardware. Turning a freeform scientific protocol into reliable robot instructions is a complex interdisciplinary problem with a mixture of context sensitive language problems, tacit knowledge embedded in domain experts, and hard optimization problems with fault intolerant and right-the-first time constraints that necessitate deterministic correctness. You will be tackling problems with few points of comparisons, deciding how to model them, shipping it, and watching real scientists use what you make. QUALIFICATIONS - 3+ years building software professionally - You have demonstrated experience taking fuzzy ambiguous problems and turning them into working shipped features independently - Comfortable jumping between languages and parts of the stack. We don't expect you to know OCaml on day one, but we do expect you to be the kind of person who'd be excited to pick it up - You have strong opinions about code, held loosely. You should want to leave the codebase better than you found it while balancing engineering momentum - You like working in-person with a small team where the line between product and engineering is blurry NICE TO HAVE - Experience or interest in domain specific languages, language design, and projectional editors - Background in the life sciences, lab automation, or software work in any domain where the user is an expert in something you're not. - Experience with OCaml, or with other ML-family languages (Haskell, F#, Rust's type system, etc.) - Experience with Next.js, Tailwind, FastAPI, NeonDB, & Zustand - Experience or interest in compilers, interpreters, heuristic algorithms and optimizing systems, especially unconventional and domain specific tools - Familiarity with LLM-in-the-loop systems: parsing, structured extraction, evals - Startup or small-team experience, especially earlier stage. LOGISTICS - Location: San Francisco, in-person at our office in SoMa. - Compensation: $140K – $180K · Offers Equity - Process: intro call → take-home project → follow up technical conversation → onsite technical interview → offer.

Requirements

Must-haves

You will be tackling problems with few points of comparisons, deciding how to model them, shipping it, and watching real scientists use what you make.
3+ years building software professionally
You have demonstrated experience taking fuzzy ambiguous problems and turning them into working shipped features independently
Comfortable jumping between languages and parts of the stack. We don't expect you to know OCaml on day one, but we do expect you to be the kind of person who'd be excited to pick it up
You have strong opinions about code, held loosely. You should want to leave the codebase better than you found it while balancing engineering momentum
You like working in-person with a small team where the line between product and engineering is blurry
Location: San Francisco, in-person at our office in SoMa.
Compensation: $140K – $180K · Offers Equity
Process: intro call → take-home project → follow up technical conversation → onsite technical interview → offer.

Nice-to-haves

Experience or interest in domain specific languages, language design, and projectional editors
Background in the life sciences, lab automation, or software work in any domain where the user is an expert in something you're not.
Experience with OCaml, or with other ML-family languages (Haskell, F#, Rust's type system, etc.)
Experience with Next.js, Tailwind, FastAPI, NeonDB, & Zustand
Experience or interest in compilers, interpreters, heuristic algorithms and optimizing systems, especially unconventional and domain specific tools
Familiarity with LLM-in-the-loop systems: parsing, structured extraction, evals
Startup or small-team experience, especially earlier stage.

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

pythontypescriptreactgomlrustnext.jsllm

Platform: ashby
Effort: quick