Founding Applied AI / Full-stack Engineer – Agentic Applications

Deffai; San Francisco, Northern (CA, KY); Hybrid; USD 150,000 - 230,000; Full time

Job summary

Deffai seeks a founding Full-stack Engineer to accelerate the delivery of AI-driven FDA regulatory solutions. You'll own core architecture and implement scalable data pipelines across the stack.

You'll collaborate closely with a small expert team to ship features, improve agent workflows, and validate models in a highly regulated domain.

Qualifications

• Familiarity with Python or TypeScript.
• Working knowledge of APIs, SQL, Git, and testing.
• Experience with LLM applications, tool calling, or RAG.
• Able to build, debug, and test applications using AI coding tools.
• Able to learn quickly and own features from idea to delivery.

Responsibilities

• Build backend and frontend features using AI coding tools.
• Debug agent workflows and improve quality, latency, and cost.
• Turn expert feedback into product improvements and evaluations.
• Fix a retrieval, citation, or tool-use issue in an existing workflow.
• Add an evaluation case, measure the improvement, and ship the fix.

Skills

Python

TypeScript

APIs

SQL

Git

Testing

LLM apps

RAG

AI tooling

Education

Bachelor's degree

Tools

PostgreSQL

AWS

FastAPI

React

Job description

Founding Applied AI / Full-stack Engineer – Agentic Applications

About the company

In the near future, drug discovery will no longer be the bottleneck in life sciences: Eli Lilly's $2B+ partnership with Insilico Medicine, Anthropic's $400M+ acquisition of Coefficient Bio, and OpenAI's partnership with Novo Nordisk. Capital and resources are flowing into drug discovery, but this doesn't address the downstream bottleneck — if there is an abundance of molecules discovered, how eligible are they for regulatory approval and commercialization?

Deffai is reimagining how drugs are approved by the FDA with cutting-edge AI purpose-built for medical devices, drugs, and therapeutics companies. We're building the world's largest FDA regulatory AI by combining FDA regulatory expertise and data from diverse sources.

Our team is made of passionate domain experts and experienced entrepreneurs across life sciences and software products. We're a mission-driven and energetic team excited to build the future of FDA regulatory approval.

We're growing quickly and looking for ambitious builders who want to tackle hard technical problems, move fast, and have real impact on how medicines are made and approved.

About the role

This is a rare opportunity for a hands-on builder to work across the full product stack, build 0 → 1 and grow with a young team. We're looking for a founding engineer who is passionate about solving challenging problems in a highly regulated industry. You don't have to know FDA or biotech, but you believe in vertical AI solutions and want to push the boundary of what is possible.

Backed by top-tier investors from applied AI and infra, with advisors from leading AI companies.

You will work closely with 2 engineers on the team, build, ship quickly and shape our product vision.

You'll work on the full production pipeline including data pipeline, database, RAG, multi-agent orchestration, eval benchmarks for internal agent performance and frontier model benchmarking.

You'll work with diverse data sources and cross-domain with experts to translate domain knowledge into structured databases.

Key responsibilities

• Build & Ship: Own architectural decisions that balance innovation, scalability, cost, and long-term maintainability in a high-stakes domain.
• Ingest at Scale: Build robust, scalable pipelines to ingest, extract, and structure heterogeneous regulatory sources into training- and retrieval-ready datasets.
• Data Quality: Ensure data quality, provenance, versioning, and governance across the corpus, with automated validation and monitoring.
• Benchmark & Evals: Build datasets, define rigorous metrics, and measure model performance across high-impact AI tasks to guide development.
• Human Annotations: Build scalable pipelines to collect structured human feedback, benchmark subjective quality, and inform model iterations.
• FDA Collaboration: Work with former FDA reviewers to accelerate their regulatory review workflows.
• Retrieval & Agents: Improve RAG, citations, agent tool use, and memory across complex documents.

Day to day, you'll

• Build backend and frontend features using AI coding tools.
• Debug agent workflows and improve quality, latency, and cost.
• Turn expert feedback into product improvements and evaluations.
• Fix a retrieval, citation, or tool-use issue in an existing workflow.
• Add an evaluation case, measure the improvement, and ship the fix.

Requirements

• Familiarity with Python or TypeScript
• Working knowledge of APIs, SQL, Git, and testing
• Experience with LLM applications, tool calling, or RAG
• Able to build, debug, and test applications using AI coding tools
• Able to learn quickly and own features from idea to delivery

Nice to have

• Experience with agent workflows, evaluations, or tracing
• Familiarity with FDA regulatory workflows or scientific literature
• Healthcare, life sciences, biomedical, or FDA-related experience
• Experience with React, FastAPI, PostgreSQL, AWS, or agent evaluation tools

Who you are

• Hands-On Builder: You love to build and ship.
• Strong opinions: You have strong opinions and product taste, and you're not afraid of pushing back.
• Real-World Experience: You've shipped things — products, pipelines, side projects — and can show us.
• Execution-Oriented: You move fast, take ownership, and focus on solving real problems.
• Clear Communicator & Team Player: You collaborate well across functions and push decisions forward.

Job details

• Location: San Francisco Bay Area (on-site) or fully remote within the US
• Work authorization: Must be authorized to work in the United States; this role does not offer visa sponsorship.
• 30min intro with CEO & cofounder
• 40min tech call where you meet our engineering team
• Short paid take-home (~5hr) for mutual evaluation