Tanay Jagadeesh
engineering @ uwaterloo | build. ship. deploy.
building AI agents and shipping real products.
about
i won 1st overall 🏆 at CXC — Canada's Largest Data and AI Hackathon, and i'm building on that momentum.
currently working on
voice agents with Wedge & Dr. Guha Roy
buildingBuilding voice-first AI agents with Wedge (YC S25) and Dr. Guha Roy for healthcare and other real-world workflows. Focus on natural conversation, context awareness, and reliable tool use so agents can assist users end-to-end.
llm analytics with sephira institute
buildingBuilding LLM-powered analytics and sentiment chatbots with Sephira Institute and Cambridge Prof. Roberto Foa. Targeting structured insights from unstructured text and deployable conversational interfaces for research and decision support.
experience
training models for an AI-native content intelligence platform for communications and marketing agencies
building llm powered analytics chatbots targetting sentiment analysis
built software for satellite systems and orbital mechanics
contributed to a rag-powered cooking assistant from the ground up
projects
real-time 1v1 coding competition platform where developers race to solve algorithmic problems. built with instant matching, live opponent status, and ai-powered hints. reached 100+ users in 2 hours after launch. fully deployed and shipping.
CORTEX
🏆 1st PlaceAwarded 1st Place Overall at Canada's largest Data Science hackathon ($1,000 prize, 300+ hackers, 80+ teams)
Full-stack 3D semantic memory system that parses ChatGPT/Claude exports, generates embeddings via Ollama, reduces dimensions with UMAP, and renders interactive Three.js visualizations. Engineered FastAPI backend with 10 REST endpoints and 7-stage ML pipeline. Implemented MCP server enabling Claude Desktop to query past conversations.
desktop productivity tool with yolo + opencv for webcam inference and habit monitoring. built in 36 hours with a team of four.
api built with yfinance and fastapi that aggregates stock news and performs sentiment analysis.
automated data pipeline for collecting, processing, and analyzing job market trends with real-time scraping and visualization.
A machine learning project that generates synthetic customer data, analyzes behavioral patterns, and predicts customer churn for a subscription-based business. Includes EDA notebooks, feature engineering, and trained models.
deep learning system for detecting forged signatures using convolutional neural networks.