Details about the role
We're automating the "human bureaucracy" layer in drug development — using LLMs to write, extract, summarize, and quality-control high-stakes documents. You’ll be designing and improving AI workflows where correctness, transparency, and iteration matter.
- Design and refine AI pipelines that use OpenAI, Anthropic, and open-source models
- Chain prompts and tools for: extraction, summarization, classification, and QC
- Build internal eval frameworks and structured test sets
- Work with domain experts to ground the outputs in regulatory logic
Expectations for the candidate
- Strong Python engineer with experience working with LLM APIs
- Experience building or customizing agent-like chains (LangChain, custom orchestration)
- Familiarity with unstructured → structured data flows (e.g. PDF → table)
- Experience evaluating model outputs (qualitative + quantitative)
- Desirable: experience with document parsing products
- Bonus: experience in clinical, regulatory, legal, or other high-precision domains
Opportunities
- You won’t be optimizing click-through rates or running AB tests on a landing page — you’ll be building real-world AI systems that touch billion-dollar regulatory decisions in pharma and biotech.
- End-to-end ownership over pipelines
- Chance to define system architecture and influence product logic
- You’re not slotting into a team — you are the team
- Work with a founder who is both technical and trusted by pharma
If you want your next 12 months to matter — and to work on systems that will be referenced in regulatory AI standards and biotech workflows — this is it.
Applying for a job
One can apply until a suitable candidate is found or until the end of the application period.