Head of AI Engineering

About the Company

AIOS is building the world's first full-stack AI doctor. Currently at $350M ARR and growing rapidly, the company serves over 150k patients monthly via its UK brand, Bolt Pharmacy. AIOS aims to become the dominant full-stack AI healthcare provider in Europe by automating clinical decision support and prescribing through advanced agentic workflows.

Responsibilities

• Build and own the AIOS Agent SDK as a strongly opinionated internal platform
• Act as the senior technical owner for agentic systems like Jesse (customer support) and Aegis (clinical support)
• Establish robust evaluation systems using deterministic checks, simulations, and model-based graders
• Manage the production learning loop by turning traces and incidents into improved architecture and models
• Ensure agent safety through authorization, validation, idempotency, and human handoff protocols
• Oversee model strategy, including selection, routing, and fine-tuning of open-weight models
• Lead the Applied AI team, growing it to approximately five exceptional engineers
• Maintain hands-on involvement by writing production code and setting technical architecture

Requirements

• 8+ years of software engineering experience with active production contribution
• Bachelor's degree in Computer Science, Machine Learning, or a related technical field
• Proven track record of building and shipping exceptional agentic systems in production
• Deep understanding of agent architecture, orchestration, retrieval, and memory
• Experience building evaluation systems for probabilistic products
• Strong systems-engineering fundamentals (APIs, distributed systems, concurrency, databases)
• Experience leading and managing small technical engineering teams
• Ability to communicate complex technical ideas to diverse stakeholders

Preferred Qualifications

• Experience building AI runtimes, SDKs, or shared infrastructure
• Background in high-stakes systems such as healthcare, finance, or insurance
• Experience fine-tuning or deploying open-weight models for specific workflows
• Expertise in long-term memory, context compression, or real-time/voice AI systems
• Strong relationships with frontier model providers