Mikalai Drabovich is an AI Engineer with 11 years of experience building high-performance ML pipelines and numerics for custom hardware, currently working on AI workloads at Quadric. He has deep expertise in LLM workload simulation, handcoded and compiler-generated kernels, microscaling numerical formats, and collective/barrier optimizations from his Sr. Staff role at d-Matrix. At AMD he implemented end-to-end training and inference on AMD-only hardware and developed TensorScope and DeepSpeed-ROCM tooling to pinpoint TFLOPs and memory bottlenecks—skills that combine systems-level profiling with model-aware optimization. His background spans computer vision on mobile devices, research into fixed-precision DNN mappings, and even launching consumer apps with millions of downloads, reflecting a rare mix of low-level hardware-software co-design and product-focused engineering. Based in Palo Alto, he brings practical experience turning numerical research into production-ready ML pipelines on alternative architectures.
10 years of coding experience
15 years of employment as a software developer
Master's Degree Radiophysics and Electronics, Master's Degree Radiophysics and Electronics at Belarusian State University
Heriot-Watt University Edinburgh Campus
Artificial Intelligence AI, Artificial Intelligence AI at Stanford University
Contributions:1 release, 22 commits, 2 PRs in 3 months
pedestriansreal-time-detectioncarsroadsreal-time
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.