At ArcFit, I built and deployed an LLM-powered Diet & Nutrition AI Agent and shipped an AI Chat Support Agent for member app-usage queries. I engineered retrieval, context-management, and database-backed workflows as a core engineering contributor.
On my SHL Assessment Recommendation Bot project, I designed and deployed a live RAG system on AWS EC2. It uses Sentence-Transformer embeddings and FAISS for semantic retrieval, with the Groq LLM API to produce contextualized recommendations.
I also built a churn prediction pipeline and deployed it through a Flask app, selecting a model that reached 93% accuracy on held-out test data. For my AI Push-Up Form Correction System, I used MediaPipe pose estimation and a Random Forest classifier to detect incorrect form with 88% accuracy.
