At Maxgen Technologies Pvt. Ltd., I worked on data analysis and machine learning pipelines, built and evaluated regression and classification models, and prepared real-world datasets. I also created Power BI and Tableau dashboards to track performance metrics and support business insights.
I built a Flask-based mental health support chatbot and benchmarked eight LLMs using semantic, lexical, and efficiency metrics. Mistral-7B achieved a BERTScore of 0.879, while DeepSeek-Chat was the most balanced overall; I also designed an intent-matching and fallback pipeline for low-confidence queries.
For my BNSS Legal Assistant, I built a RAG-based chatbot that retrieves relevant provisions from the Bharatiya Nagarik Suraksha Sanhita, 2023. I indexed 673 legal text chunks in ChromaDB and connected Llama 3.2 through Ollama to generate context-grounded answers in a Streamlit interface.
I also built a multimodal autism screening system combining questionnaire data with facial-image analysis, applying SHAP and LIME to explain predictions. My research on an AI-powered chatbot for psychometric testing was presented orally at ICACM 2026, and I won first prize at VIT Chennai’s Datathon after analyzing a bank scam dataset.
