Lead Qualcomm's Automotive AI Software Engineering Group, managing a globally distributed team and overseeing product strategy for automotive AI hardware and software solutions in ADAS and infotainment systems.Drive hardware and system innovation to enhance efficiency, performance, and power optimization in automotive AI applications, influencing design decisions to improve concurrency and reduce power consumption.Architect and implement AI software platforms and SDKs for automotive applications on Qualcomm's SoC platforms, maximizing performance through advanced optimizations and enhancing neural network architectures.Advance AI technologies, leading developments in Large Language Models (LLMs), Generative AI, and quantization techniques to improve accuracy, reduce latency, and increase efficiency in AI models.Collaborate across global teams and business units, managing cross-functional projects to integrate hardware and software solutions and driving innovation in automotive AI technologies worldwide.

Leading a team of engineers to develop QCOM next generation autonomous driving platformResponsible for architecture and design of hardware/software for camera processing (ISP), neural network acceleration, computer vision acceleration, ADAS platform SDKFocus on software functional safety (ASIL, ISO26262), design robustness, scalability, performance and virtualization across multiple OSes (QNX, Green Hill, Linux etc.)

Leading a team of engineers to develop automotive ADAS system utilizing computer vision, hardware acceleration on Qualcomm ASICs

Responsible for design, implementation and deployment of ADAS features for Qualcomm automotive chipsets on various platforms (Linux, Android, QNX, Hypervisor). - Multi-camera Surround View - Birdeye view- Surround view with 3D modeling and projection of the car and surrounding- Rear View Camera (Object detection)- Lane departure Warning system- Pedestrian detection- Driver distraction detection systemSome of the topics of research and development are lens correction, image stitching, blending, color correction, OpenGL / Open CV optimizations, Vector Extensions/Engine utilization

As a Voice Software Team Lead, responsible for architecture and design for low power audio/voice technologies in mobile space, including hardware, software, firmware:* UMTS/CDMA based technologies and system design* LTE/VoIP related codecs, bandwidth extension and features designAs a SW Product Engineer Lead, worked with multiple teams in development and deployment of EVS voice codec (3GPP TR 26.952). Its a unique and challenging multi-disciplinary technical role with some product management responsibilities:- End to end planning from design stage, development, testing and final delivery to customers- Drive discussion on spec compliance, deployment plans with both infra vendors as well as mobile vendors- Help the teams narrow down broad product level requirements to system level and then further broke them down to the each of the different subsystems (SW, HW, FW, System etc)- Technical discussions and contributions to optimize design (Static and dynamic memory footprint reduction, use multi-threading modeling to help predict MIPS utilization, lower voltage and clock footprint using dynamic clock gating, end to end timing performance for round trip delays)- Coordinate requirements and facilitate back and forth discussions to help provide feedback and get all teams on the same page including R&D, Software, Hardware, Firmware- Test planning in consultation quality assurance (QA) teams, application engineering and field test engineers- Coordinate and help design tools for voice tuning, profiling performance and for processing offline and online logs - Help coordinate the teams, drive in lab bring up and testing using daily/weekly SCRUMAs a Audio Team Code Review board member, one of two individuals in an organization of 200 engineers. The role involves enforcing best code practices and reviewing bugs identified by static code analysis tools like LINT etc to ascertain whether the bugs are real or short coming of the analysis tools.

Image processing pipe for features such as Auto White Balance, Auto Exposure and Auto FocusResearched, developed and integrated Red Eye Reduction feature in the AMD Handheld’s chipsetFace detection, smile detection and blink detection in to AMD’s driver stack. Including auto capture triggering based on different conditions (prevent blinks in picture for example).Researched and enhanced Auto Focus algorithm by integrating with face detection to accomplish clearer focus on areas of interest in a picture.Proposed feature extensions to internal OpenMax committee for general body acceptance.

Maintained and developed features of an in-house RTOS for the DSPs (Audio and Video).Inter Processor Communication driver to facilitate communication between ARM, and Audio/Video DSPs (Tensilica).Dynamic loading of images for the DSP on a per use basis to improve memory utilization.Library for IPC, cache attribute settings, RPC and other various functions for the firmware tech teams.Interesting issues such as cache coherency, random memory corruption, ASIC bring up issues etc.

Co-Founded the company providing portal solutions to business for internal intranet, product called Click-Server portal.Click-Server portal developed using Microsoft .Net Platform and as such had the distinction of being one of the first commercially available software implemented in Microsoft .Net, C# and ASP+Implemented the design utilizing object oriented database.