Senior Staff Engineer
CurrentDec 2023 - till date:Lead and developer of AISW Accuracy Evaluator Tool - Part of global AI software team, working on common AISW stack - Supporting target devices across mobile, auto, compute, IoT, and XR platforms - Leading and contributing to ML Accuracy Evaluator tool feature development - Supporting various ML models including LLMsSep-2020 - Nov 2023:Lead of Machine Learning Accuracy Team for Qualcomm's AI Cloud division. - Perform accuracy analysis on ML models & LLMs using accuracy tools and resolve accuracy issues. - Explore mixed precision, quantization, compression techniques to trade-off accuracy with speed. - Optimize ML models and tune parameters to meet accuracy targets on AI100 inference acceleratorFeb-2019 - Aug-2020:Software Development for Machine Learning tools for Qualcomm's IoT division. - Tools used for training, optimizing and evaluating the object detection model, - Severe memory constraints, light weight model used with ultra low power (< 1 mW) camera - Use cases: Retail/IoT, Video Door Bell, Low Cost / Security Camera, Interactive Toys.Feb-2018 to Jan 2019: Data Server Firmware DevelopmentSoftware and Firmware development for the Server Platform chipsets from Qualcomm Data Server.- Handled overall ownership of RAS module. (Reliability, Availability & Serviceability).- Developed RAS drivers for Dakar Mesh & DDR.- Supported of EMAC driver for new chipset. - Supported IMC firmware (Integrated Management Controller).