AI/ML Intern
Information Technology graduate with hands-on experience building RAG chatbots, hybrid retrieval systems, sentiment analysis models, and OCR-to-LLM document processing pipelines.
Vyapar Taxone
November 2025 - May 2026 · Onsite
Engineered automated data extraction pipelines using pdfplumber for complex banking PDFs; automated end-to-end document ingestion workflows; handled low-resolution scans and multi-column layouts; developed OCR-to-LLM pipelines that improved extraction accuracy from 40% to 80%; created an AI-based scripting agent to generate parsing logic for new bank statement formats.
Bachelor of Science, Information Technology
2022-25
CGPA: 6.96
Built a ChatGPT-like system for answering questions from PDFs using hybrid retrieval with FAISS and BM25 combined through Reciprocal Rank Fusion. Developed a FastAPI backend with JWT authentication, multi-session chat, and PDF indexing using Python, LangChain, LangGraph, Whoosh, and SQLAlchemy. Improved retrieval coverage from 50% to 100%.
Developed a RAG-based chatbot for technical documentation with document grounding to reduce hallucinations. Implemented vector search using ChromaDB and HuggingFace embeddings, plus a Streamlit UI with JWT authentication and persistent PostgreSQL chat storage.
Built a machine learning model to classify sentiment from tweets and comments. Compared Logistic Regression, SVM, and Naive Bayes using TF-IDF features, with Logistic Regression achieving the best performance. Performed tokenization, stopword removal, and vectorization.