AI Product Manager · B.Tech CS + AI @ IIIT-Delhi
Building production AI systems — from voice agents to quantitative trading platforms
|
market-intelligence-platform (private) Enterprise-grade Indian Market Intelligence Platform — Agentic RAG, 14-metric fundamental scoring, 131K-record backtesting, and smart portfolio rebalancing |
stock-prediction-engine (private) ML-powered stock prediction with 350+ technical indicators and LSTM models for Indian equities — BUY/SELL signal generation |
ai-caller (private)
Enterprise Voice AI platform — 15 microservices, 3 deployment approaches (Cloud/Local/PersonaPlex), <700ms voice-to-voice latency, 36K+ call transcripts
|
RAG-powered system for downloading, indexing, and analyzing cement industry conference call transcripts with LangChain and ChromaDB |
Intelligent chatbot using LangGraph, Google Gemini, and persistent memory |
RAG-based chatbot for tech transfer knowledge base |
|
High-performance FastAPI REST API for image classification using ensemble deep learning (CLIP, BLIP, DETR) |
Real-time object detection with YOLOv8, deployed via Flask API |
XGBoost-based coupon tier classifier predicting optimal High/Mid/Low discount levels based on user behavior and engagement patterns |
|
Speech-to-text + sentiment analysis API using Whisper and transformers |
Text summarization tool using NLP techniques (extractive + abstractive) |
labs-financial-data-pipelines (private)
Automated financial data collection — broker emails, recommendations scraping, factsheet downloads, and portfolio analysis
| Project | What it does |
|---|---|
| propensity | Propensity scoring model for user conversion prediction |
| rocket-flight-controller | Rocket flight simulation and PID controller implementation |
| Article-Summary-Insights | AI-powered financial article analysis system with multi-agent architecture for insights extraction and investment scoring |
| cohortx-task-2 | CohortX Challenge Task 2 - eligibility criteria to semantic triples. MMR retrieval-augmented few-shot extraction. 3rd of 9 validated teams. |
| finance-rag-system | Enterprise-grade, finance-focused Agentic RAG system with multi-agent architecture, GraphRAG, and zero-mismatch numerical verification |
AI/ML: PyTorch, Transformers, LangChain, LangGraph, LSTM, XGBoost, CLIP, YOLO, Whisper, RAG, Reinforcement Learning
Quant: 350+ Technical Indicators, 14-Metric Fundamental Scoring, Backtesting, LSTM Prediction, Smart Rebalancing
Backend: FastAPI, Flask, Node.js, TypeScript, PostgreSQL, Redis, Neo4j, Docker, AWS SQS
Frontend: React, Vite, Streamlit, Chart.js, Tailwind CSS
- B.Tech Computer Science + AI @ IIIT-Delhi (Minor in Entrepreneurship)
- Former Research Intern @ SBI Labs (Survival Analysis) and ECE Labs (NLP/LLMs)
- Technical Secretary — Student Council, IIIT-Delhi
- Founder — CyFuse (Tech Club)