Summary
Indrashish Basu is a Member of Technical Staff at AMD with nine years of focused experience in GPU performance optimization, machine learning workloads, and scalable infrastructure. He has driven measurable gains—boosting ML application efficiency from 67% to 94% on MI300A/MI300X platforms and cutting workload processing from months to hours through clustering and profiling innovations. A pragmatic engineer fluent in Python, C++, and shell scripting, he built client-server infrastructures, database migration tooling, and profiling frameworks that reduced manual analysis from days to minutes. He combines low-level hardware counter analysis and power/performance tuning with higher-level systems design, enabling both chip bring-up and production ML training (e.g., Llama on MI300X). Based in California, he mentors cross-functional teams and collaborates closely with architects to translate performance insights into architecture-level improvements. Colleagues rely on him for turning complex performance bottlenecks into repeatable, scalable solutions that accelerate time-to-insight.
9 years of coding experience
11 years of employment as a software developer
Science, Science at South Point High School
Bachelor of Technology (BTech), Electrical Engineering, Bachelor of Technology (BTech), Electrical Engineering at West Bengal University of Technology
Masters of Science, Electrical and Computer Engineering, Masters of Science, Electrical and Computer Engineering at University of Florida
English, Bengali, Hindi