abhilash User

I’m a final-year B.E. CSE (AI/ML) student with hands-on experience building machine learning, NLP, and deep learning systems. I enjoy designing end-to-end ML pipelines—fine-tuning transformer models and turning them into reliable, usable products.

In the Tata Group — Data Analytics Simulation (Forage Virtual Program), I conducted EDA on 5,000+ customer records, designed a no-code predictive modeling solution that reduced manual risk analysis time by ~30%, and built an AI-driven collections strategy to prioritize high-risk customers. I also used GenAI tools to enhance data quality assessment and automate analytical workflows.

I’ve built agentic RAG and deployed it as an application with FastAPI and Streamlit, including hybrid retrieval (dense + sparse scoring), multi-hop query planning, and a robust fallback/failover architecture integrating OpenAI, Groq, OpenRouter, Tavily, and Serper. Alongside that, I’ve developed a fine-tuned BERT fake review detection system (94% accuracy) and a network anomaly detection model with a full preprocessing-to-deployment pipeline via Flask APIs.