At Avench Systems, I designed and built a tool that generates unit tests for embedded C firmware using a locally deployed LLM. Its context pipeline uses Clang/libclang AST analysis to assemble function-specific context.
I engineered a generate-verify-repair loop that feeds GCC and linker diagnostics back into bounded test repair, and validated it against an STM32/HAL codebase. I also developed an AI-assisted MISRA C compliance framework combining local LLMs, PC-Lint, AST-based parsing, GCC validation, and automated rollback.
At Bangalore University, I taught graduate-level Big Data and Theory of Computation coursework and led Machine Learning lab sessions. My projects include a Databricks churn-prediction pipeline and a meeting-insights system using Whisper and the Gemini API.
