Apr 2024 – Aug 2025 · 1 yr 4 mos · Sunnyvale, California, United States · Remote

• AI & Software Engineer
• Launched AI-powered suggestions in EkLine GitHub Actions, automating PR comment fixes via Reviewdog and Vale diagnostics.
• Optimized EkLine CLI to analyze only changed files, reducing CI run times by up to 90% and lowering costs.
• Designed and implemented the Quality Metrics and Scores systems using Vale and AI to quantify documentation quality.
• Proposed and Implemented Bitbucket Integration, enabling seamless multi-platform workflows.
• Resolved a critical bug in EkLine Actions related to non-ASCII character handling.
• Proposed and Implemented the “Improve Writing in Current File” command to the EkLine VS Code extension, enhancing in-editor productivity.
• Upgraded Vale for EkLine CLI and Cloud Functions to support new file structures and prevent future compatibility issues.
• Built and deployed the Docs Assistant from scratch using chunking, vector embeddings, and vector search (RAG) for real-time, context-aware documentation also integrated with a custom UI based on a forked Chainlit build.
• Designed and Implemented organization-aware context in Docs Assistant to improve the relevance of generated content.
• Developed a scalable event metrics system, enabling detailed analytics and usage-based billing.
• Implemented a system to generate documentation PRs from Linear tickets by integrating Firestore RTDB with Cloud Run, using the Docs Assistant's FastAPI backend interface.
• Integrated Google Docs, Slack, and Linear into Docs Assistant using agentic tool calling via Composio, expanding its platform capabilities.
• Built LLM evaluators to detect regressions and ensure high-quality output with each release of Docs Assistant.
• Implemented multi-tenant support in EkLine CLI, removing synchronous execution bottlenecks.
• Proposed and developed the EkLine Model Context Protocol (MCP) to provide seamless integration within users' preferred IDEs.