No sponsorship (H/F1, EAD, OPT) or C2C. Principal applicants only
DIRECT APPLICATION: https://www.careers-page.com/implaion/job/X96Y796R
We're building the intelligence layer for how the life sciences industry makes its most consequential decisions — drug acquisitions, licensing deals, and investment calls worth hundreds of millions of dollars.
Our platform uses LLMs and RAG to rapidly synthesize complex scientific and financial documents into structured, trustworthy assessments for pharma executives and investors who need to move fast without sacrificing rigor. The AI isn't a feature — it's the product.
The role
You'll own the full AI stack. That means designing and shipping production RAG pipelines, architecting how LLMs interact with heterogeneous scientific and regulatory document corpora, building the evaluation frameworks that keep outputs trustworthy, and maintaining the MLOps layer that keeps everything running. You'll work directly with domain scientists and senior leadership — no hand-offs, no layers.
This is a 12-person company at Series A. You won't execute tickets handed down from a research team. You'll make the architectural calls.
What we're looking for
• 4+ years of ML/AI engineering with 2+ years specifically in LLM applications and RAG systems — not prototypes, production
• Deep hands-on experience with LLM application frameworks (LangChain, LlamaIndex, or equivalent) and the full retrieval stack: chunking strategies, embedding models, vector stores, re-ranking
• Python as your primary production language; strong Git, CI/CD, and MLOps practices
• Cloud ML deployment experience (AWS preferred) with Docker and Kubernetes
• Track record of owning systems end-to-end at a startup or small team — you've debugged failure modes in production, not just built demos
• Bonus: experience with scientific, clinical, or regulatory document processing; knowledge graph construction (Neo4j or equivalent); biopharma or health tech domain exposure
What makes this unusual The users of this system are pharma executives and investment analysts who will immediately recognize a hallucination or a low-confidence retrieval. The technical bar for trustworthy AI output here is higher than most consumer applications. If that challenge energizes you rather than concerns you, we'd like to talk.
Competitive compensation. Fully remote (US).
Apply tot his job
Apply To this Job