Archit Jain

Built and deployed a multi-stage restaurant ranking system for Zomato Dining using L1/L2 Gradient Boosted Tree rankers and a downstream optimization function, driving a 60% improvement in contribution margin per transaction. Owns the ML charter for Dining discovery and content, leading a team of 3 engineers and increasing discovery-led transactions from 4.5% to 11%.

Led a team of 5 to build a GenAI-powered AI Companion and personalized notifications, designing a scalable system using Kafka and distributed task workers. Built real-time and batch personalized game recommendations, driving up to 6.8% revenue uplift, and improved daily active-user retention from 50% to 77% through end-to-end analytics.

Built a production car damage detection system, reducing car NoGo from 30% to 5% and developing object detection models to reduce human errors. Developed LLM-based business applications, speech-to-text systems, a foundational car pricing model, and an efficient number-plate masking system achieving 35 FPS.

Developed deep learning and computer vision models for video analytics, including action/explosion classification and object detection using Keras, Feature Pyramid Networks, and Faster R-CNN. Optimized scene-change and soft-cut detection algorithms for VOD/OTT content, achieving a 4.26% accuracy improvement.

Pursued a Bachelor of Engineering in Instrumentation and Control from Netaji Subhas Institute of Technology, Delhi University, from 2015 to 2019.