At Handshake AI Fellowship, I design and evaluate challenging command-line tasks that benchmark AI agents on code execution, debugging, system exploration, and multi-step workflows. I analyze agent failures and provide structured technical feedback to improve the reliability of AI agents and LLM-based systems.
At Eskillveda Edtech Private Limited, I architected a multimodal generative AI platform integrating Vision, Audio, TTS, and Translation services. I built LangGraph workflows with OCR fallbacks and an asynchronous FastAPI backend, then containerized and deployed the services to the cloud.
As an AI Engineer Intern at Eskillveda Edtech Private Limited, I developed a full-stack AI music analysis platform and built REST APIs for audio source separation, lyric transcription, and BPM and key detection. In my rainfall forecasting project, I combined deep learning models with an XGBoost meta-learner, achieving R2 = 0.98 and RMSE = 1.55 mm on held-out test data.
