SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Sail Research is building foundational infrastructure for agentic AI systems. This Member of Technical Staff role focuses on agent engineering—moving beyond inference to design and build complete systems that turn billions of tokens into high-quality outputs.
Key responsibilities include:
- Designing custom evaluation frameworks for multi-turn, massively parallel agent systems
- Building agent harnesses to improve performance of open-source models (Deepseek, Qwen, Llama)
- Collaborating on agent/harness codesign similar to Claude Code's approach
- Automating prompt optimization techniques using frameworks like DSPy
- Contributing to core infrastructure that enables effective background agents
You'll work with a small, highly technical team in Sail's downtown San Francisco office. The role requires deep expertise in AI agent development and familiarity with open-source model ecosystems. This is a hands-on engineering position where you'll tackle some of the most ambitious technical challenges in AI infrastructure.
The interview process is designed to be respectful of your time: initial conversation with the CEO, technical discussion with the CTO, an in-person interview day at the SF office including a 3-4 hour work simulation (objectively scored, with immediate feedback), and a presentation of your process and results. AI assistance is encouraged during the technical assessment.
Sail invests heavily in team experience—beautiful office in downtown SF, all meals provided, Studio Display monitors, and a focus on tools and infrastructure that save time and energy.
