At Internova Labs, I developed an AI-powered resume screening and candidate evaluation system using Python, NLP, LLMs, embeddings, and ChromaDB. It matched skills semantically and through LLM evaluation, with results presented in a Streamlit dashboard.
I also built Nova, a voice assistant integrating Faster-Whisper speech recognition, local LLMs through Ollama, text-to-speech, and conversation memory.
At Micropro Software Solutions, I collaborated with the data team to collect, clean, and preprocess datasets, and supported the training pipeline of an AI product.
In my projects, I built a RAG-based resume screening pipeline and an AQI prediction system that benchmarked five models, with XGBoost performing best. I also built an offline desktop voice assistant with tool routing and persistent JSON-based memory.
