At Samsung Innovat, I engineered a retail demand forecasting pipeline across 100K+ rows, improving accuracy by 38.1% over the seasonal naive baseline. I also benchmarked regression models and developed a Streamlit dashboard for forecast monitoring.
At Infosys, I developed an AI-powered Content Marketing Optimizer using Transformers, Groq, and Gemini, achieving approximately 85% sentiment classification accuracy. I also integrated Slack and Google Sheets APIs to automate campaign reporting.
I built an Enterprise AI Assistant with document Q&A using RAG and FAISS, natural language-to-SQL generation, and PDF report generation. The application uses JWT authentication and RBAC, and I containerized it with Docker and implemented GitHub Actions CI/CD.
For an airline booking automation project, I developed a 64-test Selenium and pytest suite that reduced regression execution time by approximately 62%. I also built a keystroke dynamics authentication system using a PyTorch LSTM model, FastAPI, and PostgreSQL.

