Jeevan Bhooi
Senior Staff Software Engineer @高通
Signup · Get unlimited contacts
WORK HISTORY
Senior Staff Software Engineer @高通
Markham, ON, CA
GenAI tech lead for automotive, driving the successful delivery of multiple high impact projects, customer proof of concepts and initiatives on Android, Linux and QNX.o Lead for initiatives in Large Language Models (LLMs), Generative AI, and quantization techniques to increase model efficiency, reduce latency, and improve accuracy.o Cross team collaboration to onboard multiple LLMs, VLMs and multimodal models from the Llama, Qwen, Phi and other family of models to run quantized workflows on Qualcomm SoCs platforms.o Used AI Model Efficiency Toolkit (AIMET)- an open-source library, on Nvidia A100 tensor core GPUs for optimizing trained neural network models.o Boosted decode performance with different speculative decoding methods – SPD, SSD, Eaglet. Prepared draft models for speculative decoding using multiple A100 GPUs.o Prepared Jupyter notebooks for customers to use and deploy GenAI solutions on edge.o Setting up benchmarking of LLMs and VLMs with MMLU, benchmarks and validating performance with different dataset (XLAM, LongBench, LVEval).Contributed towards the design and development of AI software platforms and SDKs for automotive use cases on Qualcomm’s SoC platforms, delivering advanced performance through cutting-edge optimizations and neural network architecture enhancements. Qualcomm AI runtime (QAIRT):o Added support for different functionalities in QNN like asynchronous execution, device manager for managing the lifetime of and interacting with multiple DSP devices and cores, interface for offline simulation metrics for different SoCs, interface for users to iteratively profile different compiler options.o GenAI use cases- Optimizations for weight sharing and spill-fill buffer sharing between graphs- Prefill time optimization using multiple DSPs- Currently involved in design discussions for various batching solutions for efficient utilization of available DSPs.
EDUCATION
Little Angles High School
High School, High School/Secondary Diplomas and Certificates
Veermata Jijabai Technological Institute (VJTI)
Bachelor's Degree, Electronics
Vivekanand Education Society's Institute Of Technology
Technical Diploma, Electrical, Electronics and Communications Engineering
ABOUT JEEVAN BHOOI
Have more than cumulative 17 years of experience with working in Automotive AI, real time video place-shifting, client/server environment and various DTV standards.• Have more than 5 years of experience working on automotive AI.• Extensive experience in C and C++, Linux based firmware development for embedded real-time applications. Extensive experience working in embedded micro-controller based system software design, multithreaded embedded real-time systems, device drivers, communication protocols, client/server systems.• Working knowledge of networking concepts.• Have experience in leading development of firmware components. Utilized opportunities available to showcase ownership, quality & timely completion of deliverables.• Experience in middleware modules for DTV & Set top boxes.• Adept with problem solving, issues at system as well as detailed design level.• Exposure to using QT framework.• Proficient in the use of Tableau - interactive data visualization tool.• Exposure to SQL data analysis for debugging field issues.• Proficient in using SVN, Git and ClearCase for version control.• Exposure to HLS and MPEG DASH streaming protocol.• Proficient in use of online tools - JIRA, Bitbucket, Crucible/Fisheye and Coverity,• Skilled in Agile (Scaled Agile Framework) software development methodology.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.