SUHAS S.

- Worked in the pioneering integration of transformer models with deep neural network (DNN) compilers, enhancing neural network optimization and deployment, as part of a collaborative research team at SPORT Lab. - Innovated in the domain of DNN accelerators by participating in the design of specialized hardware, focusing on parallelism, memory hierarchy, and energy efficiency to optimize DNN computations. - Authored and maintained over 2000 lines of high-quality Python code and 500+ lines of C++ code, contributing to the development of a proprietary Reduced-Memory-Access Inference Compiler, elevating system performance and efficiency. - Streamlined the compiler architecture by working on graph optimization, code generation, and effectively mapping neural network models to the accelerator architecture, resulting in a 25% enhancement in compiler speed. - Conducted rigorous testing protocols to validate the performance, accuracy, and efficiency of DNN accelerators and compilers, ensuring reliability and performance in line with industry standards. - Pioneered the development of algorithms optimized for execution on resource-constrained platforms, improving computational efficiency while reducing memory usage. - Engineered and implemented custom operators to extend the functionality of deep learning models, including the conversion of high-level models into intermediate representations for efficient processing. - Orchestrated the integration of the TVM deep learning compiler into the lab's existing workflow, laying the foundation for a robust hybrid compiler system. - Accelerated deployment times by 30% through scripting tools in Python for generating HDL files, enabling seamless deployment of DNN models across varied hardware configurations. - Enhanced project efficiency by authoring detailed documentation of design, implementation, and testing procedures, ensuring clarity and maintainability of the codebase.

Performed the following functions -overlook operations, services , planning, sales ,maintenance or research and development to affect operational efficiency -drive key manufacturing initiatives in cost reduction , quality improvement and timely delivery of goods -involved in making sure production output meets product specifications ,quality specs, customer due dates -worked with materials management to help drive key initiatives relating to safety stock position, cost reduction, supplier quality improvement , inventory management -conduct daily follow up on outstanding quotes, conduct outside sales calls, marketing new featured products, follow up on leads to generate new business -research customer complaints related to aspects of service including order accuracy, material quality, damages , later and incomplete delivery