Archishman Singh

I am a BITS Pilani student studying electrical engineering with advanced mathematics training, focused on machine learning and deployed systems. I combine rigorous fundamentals with a drive to ship production-ready ML products.

I built InvestiSense AI, a finance Q&A system using RAG (Sentence Transformers + Pinecone) over 2K+ SEC filings, integrated GPT-OSS-120B (ChatGroq) via LangChain and real-time yfinance data, and deployed a FastAPI backend with a Streamlit UI supporting three user personas.

I developed DocuMentor by fine-tuning CodeBERT for comment-code alignment (75% accuracy on 100K+ Python samples) and implemented few-shot prompting for docstring generation. I also built a credit-card fraud detection pipeline (50K+ transactions) achieving 0.98 ROC-AUC using SMOTE-TOMEK, Optuna-tuned models, and a custom credit scoring algorithm.

As a Teaching Assistant, I supported teaching software design and logic to 100+ students, and I lead as captain of my college kabaddi team. I am collaborative, production-focused, and eager to join teams that value practical ML deployment and measurable product impact.