Prince Kumar

Demonstrated expertise at Indian Space Research Organisation by optimizing modeling pipelines with Python and Linux, enhancing reproducibility by 40%. Skilled in problem-solving and technical communication, delivering impactful presentations and detailed reports to streamline complex data analytics workflows.

• AIR 16637, Category(4157), Top 1%, JEE Advanced 2023, 2023
• Gold Medal, Team lead, IITI Summer of Code, 2024
• 12th (Overall 2nd), Inter IIT Tech Meet 14.0 - Pathway, 2025
• Smart India Hackathon, National Finalist, Team lead, 2024
• Pentathon National Finalist (Waitlisted), India's 2nd National-Level Penetration Testing Competition, 2025
• Rank: 4058, Leetcode Biweekly Contest 160, (300+ questions solved in Leetcode), 2025

• Video Conferencing App with Emotion Detection: Expresso, Project Lead, IITIsoc, IIT Indore, 06/24, 08/24, GitHub, Architected a real-time Computer Vision pipeline for live facial emotion recognition, benchmarking various Deep Learning architectures to balance classification accuracy with inference speed., Optimized the model inference engine and integrated it with a Node.js SFU and WebRTC, successfully reducing processing latency by 90% to enable seamless analysis across 20 concurrent video streams., Deployed the end-to-end ML serving and application stack (Django, Next.js) on an AWS VPS using Docker and CI/CD pipelines, ensuring a scalable, load-balanced environment for 1000+ users.
• Pneumonia Detection, Personal Project, 05/24, 06/24, GitHub, Built two CNN models to classify chest X-rays using the Kaggle dataset for pneumonia detection., Achieved 91% accuracy with a custom CNN and 96% using VGG16, showcasing gains via transfer learning., Explored deep learning in medical imaging, improving accuracy and classification speed for real-world use.
• Real-Time AI Copilot for Tactical Hedge Fund Managers, Team 76, Inter IIT Tech Meet 14.0, Month Year, Month Year, GitHub, Developed a zero-latency multi-agent system using Pathway and MCP, integrating a CRAG pipeline on SEC filings for verifiable, streaming-native market intelligence., Engineered a Trigger Agent hot-swapping Chronos-2/SARIMAX for live forecasting, and a deterministic Portfolio Agent utilizing Mean-Variance/CVaR solvers for strict risk compliance., Designed an STVGP-based Strategy Agent achieving a 72.4% predictive hit rate, deploying the full data stack via Docker, Kubernetes, Kafka, and AWS S3/PostgreSQL.

• Web Development Lead and Co-Head, Google Developer Group (GDG), IIT Indore, 11/24, 11/25
• Data Analysis Head, Astronomy Club, IIT Indore, 11/24, Present
• Member, Gymkhana Web Team, IIT Indore, 08/24, Present

• Computer Programming
• Data structure and Algorithms
• Blockchain