Senior AI Engineer

Weave•San Francisco, CA

3h•$175,000 - $215,000•Hybrid

About The Position

We're seeking a hands-on Senior AI Engineer who can bridge the gap between data and biomedical expertise while delivering robust backend code. Reporting to the Head of Applied AI, your primary focus will be building the LLM-facing components of our backend service, while ensuring the AI models consistently produce accurate, grounded, useful, and reliable outputs under real-world load. The ideal candidate thrives in the full development cycle: taking concepts from initial conception through rigorous experimentation and testing, all the way to production deployment. You won't get lost in low-ROI side quests, but instead will maintain velocity with unwavering focus on the core value of your tasks. We value pragmatic problem-solving over flashy technology and buzz words, robustly tested solutions over demoware. We're building solutions that actually work—simple, thoroughly tested, and rock-solid reliable. If you believe the best technology is the one that consistently delivers results, you'll fit right in.

Requirements

Direct experience building, shipping and operating meaningful LLM-powered products or workflows in production
Scientific or quantitative background in machine learning, statistics, computer science, biology, chemistry, or a related field; experience in biotech, pharmaceuticals, or another life-sciences setting is preferred
Humble, hungry, and smart attitude
Ability to move quickly while maintaining a high bar for correctness and reliability
Clear technical judgment: able to explain trade-offs, prioritize high-value work, and stop low-ROI experimentation

Nice To Haves

History of working at early-stage, product-based startups (<100 people)
Experience developing software in regulated or high-accountability environments

Responsibilities

Writing production-ready backend code for LLM-facing components of our backend service
Building automated evaluation systems for LLM-dependent code, including regression datasets, quality metrics, failure taxonomies, and release gates
Designing and validating LLM-based solutions (prompts, context-construction pipelines, tool-calling workflows, structured outputs, and retrieval systems) and proving their effectiveness with real data and meaningful metrics
Measuring and improving LLM solutions quality, grounding, consistency, latency, token usage, and cost