At Tata Elxsi, I develop AI-enabled solutions for real-time computer vision workflows using Python and C++. My work improved pipeline scalability and perception accuracy by 15%.
I worked on object detection and depth-estimation workflows for real-time vehicle distance monitoring, achieving over 90% tracking accuracy using YOLOv8. I also optimized inference pipelines, reducing inference time by 25%.
At Tata Elxsi, I created automation and dataset-preprocessing workflows that reduced manual data preparation time by 30%. I also used Coverity and CodeSonar to resolve critical MISRA C++ defects in safety-critical software.
In my Financial Document Q&A RAG Chatbot project, I built a system for answering questions about complex financial documents such as 10-K reports. It uses FAISS for semantic retrieval and provides LLM responses grounded in retrieved source context.
