Eashan Sinha

Eashan Sinha is an engineer and founder working on AI-assisted software development and the practical design of coding agents. His focus is making these systems responsive to a developer’s intentions, grounded in the actual codebase, and accountable to human judgment.

Sinha earned bachelor’s and master’s degrees in computer science from Georgia Tech, worked as an engineer at TikTok, and founded a generative AI company accepted into Y Combinator. In 2025, he was a deployed engineer at Windsurf, helping developers apply agentic coding tools to real engineering workflows.

• Developer intent as essential context. Effective agents must understand recent edits, terminal activity, relevant files, and evolving goals. Those signals help Windsurf’s Cascade and Tab anticipate useful next steps instead of treating every request as an isolated prompt.

• Codebase understanding beyond simple retrieval. Sinha favors combining repository context, developer activity, multiple retrieval methods, and tools to ground agent behavior in the software being changed.

• Structured autonomy for coding agents. Developer-defined rules, reusable workflows, persistent preferences, and Model Context Protocol integrations make multistep agents more predictable; simultaneous sessions can support separate strands of work.

• Human-agent collaboration with a definition of done. His engineering workflow begins by exploring the codebase and specifying success, then proceeds through collaborative planning, implementation, testing, and review. Developers remain responsible for correcting mistakes and verifying that the finished code actually works.