At Bharati Defence and Infrastructure Limited, I’m developing an end-to-end AI document-generation system using Retrieval-Augmented Generation, LLM-based agents, and automated workflows. I’m also building AI automation workflows with n8n and conducting applied R&D for internal machine learning and computer vision use cases.
At Infosys, I developed and evaluated supervised machine learning models for classification and image segmentation, achieving 88–92% accuracy on structured datasets. I worked across preprocessing, feature engineering, model optimization, and evaluation using Scikit-learn, Pandas, and NumPy.
On my real-time multimodal computer vision project, I designed and deployed a system for face and gait recognition across synchronized RTSP camera streams. It achieved 95%+ identification accuracy, and score-level fusion with temporal voting reduced false positives by 40–60%.
I also built a SQL-based retrieval pipeline for ArcFace embeddings and an async FastAPI and NiceGUI dashboard, containerized with Docker. My background includes a B.E. in Computer Engineering with Honours in Data Science from ST. JOHN COLLEGE OF ENGINEERING & MANAGEMENT.
