SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Greptile is an AI code reviewer that analyzes pull requests with full codebase context to catch bugs and anti-patterns. The company reviews 5 billion lines of code monthly for 22,000+ customers spanning YC startups, big tech, finance, healthcare, and defense. Greptile has scaled to eight-figure ARR with $30M in funding from Benchmark, Y Combinator, and other top investors.
As a Research Engineer, you'll experiment with and apply cutting-edge advances in AI agents and language models to enhance product performance and capabilities. Key responsibilities include designing and building multi-agent systems, harnesses, models, and evaluation frameworks that power code validation. You'll study emerging multi-agent architectures, prototype and evaluate novel code review workflows, and collaborate with the team to integrate successful prototypes into production systems. You'll stay current with the latest research across large language models, information retrieval, and developer tooling.
The role tackles exciting technical challenges: building agents that learn coding standards through osmosis like new hires do; identifying per-customer PR feedback preferences using sample-efficient reinforcement learning to improve signal-to-noise; and autonomously deploying feature branches with AI-generated end-to-end tests to detect bugs through parallel agent-driven testing.
The team is small and talent-dense, with members who have scaled critical functions at Stripe, Google, Figma, and similar companies. Everyone wears multiple hats, and day-to-day work spans multiple areas. The company values shipping products that real users rely on and thrives in the chaos of high growth.
