Summary
Alexander Hoffman is a software engineer with 8 years of experience specializing in deep learning systems and hardware-aware model optimization, currently enabling high-throughput foundation model training on AWS Trainium accelerators. He blends hands-on PyTorch/XLA work with distributed training strategy, roofline analysis, and root-cause debugging across ML framework, compiler, and runtime layers to squeeze maximum performance and correctness out of multimodal models. Prior roles include developing automated model compression tooling and research-grade pruning algorithms, giving him a strong bridge between research and production. His background in electrical engineering and robotics informs a practical hardware-aware approach—he’s comfortable from PCB and firmware design to profiling mixed-precision training at scale. Colleagues rely on him for design reviews, mentorship, and turning performance profiles into concrete optimizations.
8 years of coding experience
5 years of employment as a software developer
B.S., Electrical Engineering, B.S., Electrical Engineering at University of Washington
M.Sc., Electrical Engineering, M.Sc., Electrical Engineering at McGill University
English, French