←  Back to Home

Projects

Things I've built

๐Ÿค–
AI / NLP

PDF Answering AI

AIES, IIT Roorkee ยท May โ€“ Jun 2024

  • Built a document Q&A system using Dense Passage Retrieval (DPR) and a fine-tuned BERT model for accurate answer extraction.
  • Developed a clean web interface using Flask, HTML, and CSS for seamless PDF upload and query interaction.
  • Achieved fast, context-aware response retrieval on long-form academic PDFs.
Python DPR BERT Flask HuggingFace HTML/CSS
โšก
EV / IoT

Hardware-in-the-Loop EV Charging

EED, IIT Roorkee ยท Jan โ€“ Apr 2025

  • Implemented an OCPP 1.6 backend using Node.js and WebSockets for real-time communication between EV supply equipment (EVSE).
  • Built Python-based charging optimization using Scipy & Matplotlib for dynamic load scheduling strategies.
  • Designed a live React dashboard to monitor active EV charging sessions, energy flow, and alerts.
Node.js WebSockets React Python Scipy OCPP 1.6 Matplotlib
๐Ÿง 
Generative AI

AI Answer Evaluation System

PwC India ยท May โ€“ Jul 2025

  • Built a RAG-based pipeline using LangChain + ChromaDB for intelligent document retrieval and answer scoring.
  • Integrated Groq LLM for low-latency, context-aware evaluation of subjective answers.
  • Developed a Streamlit UI enabling stakeholders to upload documents, pose queries, and view AI-evaluated results interactively.
RAG LangChain ChromaDB Groq LLM HuggingFace Streamlit
๐Ÿ“ˆ
Finance / ML

Stock Sentiment Analysis using Machine Learning

Finance Club, IIT Roorkee ยท January 2024 - February 2024

  • Built sentiment-driven trading bot achieving 66.67% win ratio across 3 executed trades.
  • Applied VADER NLP on financial news to generate buy/sell signals for TSLA and META.
  • Simulated portfolio strategy revealing Sharpe Ratio of -6.47 and max drawdown of $18,194.
  • Scraped and processed stock news headlines using BeautifulSoup4, Pandas, and NLTK VADER.
Python 3.8+ VADER (NLTK) yFinance BeautifulSoup4 Pandas NumPy Plotly Matplotlib Seaborn
๐ŸŒ
Coming Soon

More Projects in Progress

Open to Collaborations

Currently exploring projects in agentic AI workflows, quantitative finance ML, and embedded IoT systems. Reach out if you'd like to build something together!
AI Agents Quant Finance Embedded Systems