Working on Generative AI model for awesome snapchat lens which can understand and generate personalized images and creative snaps.

– Built a large-scale video analytics pipeline to process data from 60K+ CCTV cameras for activity recognition and operational insights.

– Developed multimodal video understanding models (vision-language models + CV pipelines) achieving 80%+ accuracy on real-world surveillance tasks.

– Implemented end-to-end computer vision systems for activity recognition, behavior analysis, tracking, and re-identification.

– Reduced operational costs by 10% by deploying ML-driven efficiency monitoring across store environments.

– Engineered a low-latency backend system for real-time sports betting with ∼10ms response times under high concurrency.

– Built and deployed backend services using Node.js, Express, MongoDB, Redis, and Socket.IO on Linux-based infrastructure.

– Designed a secure, scalable API framework for managing tokens and cryptocurrency transactions using Ethereum.

• Made production line management system which handled ₹50,00,000+/month in productions

• Created web application for tax practitioner in India which is currently used to file 750+ GST returns monthly.

• Made production management system for a manufacturing company which has been in use since 2018 have collected 10,000s of invoices and payments.

– Processed $1M+ in transactions with improved system efficiency by 10%.

– Architected and deployed microservices on AWS + Kubernetes, achieving 99.99% uptime.

– Made distributed systems with MERN, Docker, Redis, RabbitMQ, implemented asymmetric encryption for secure transactions.