BHAVESH DINESH PATHAK

Architected a Prompt-to-Game Logic AI agent using LangChain & Python, automating game-design workflows from natural-language prompts and reducing manual effort by ~65% across 3 game genres. Built 5+ LLM automation pipelines with LangGraph & N8N; integrated FAISS vector-search into RAG workflows improving retrieval relevance by ~40% over keyword-based methods. Prototyped 2 client-facing AI features end-to-end — from prompt engineering to FastAPI deployment.

Trained 3 ML models (Random Forest, Decision Tree, Logistic Regression) on a 10K-row energy dataset; best model hit ~82% accuracy on anomaly prediction. Developed an IBM Watson Assistant chatbot for energy-saving tips; produced 8+ Seaborn dashboards.