Summary
Blaine Rister is a senior software engineer in the San Francisco Bay Area focused on compiler and ML infrastructure for custom accelerators, currently on Meta’s MTIA team building the graph compiler that connects PyTorch to an LLVM backend. He brings a rare blend of deep research and product experience — a Stanford PhD and years of lab work in medical imaging paired with hands-on systems shipping at startups and industry leaders like SambaNova and NVIDIA. At Femtosense he led a compiler team, ported a Python compiler to a custom MLIR dialect, and built deployment tooling and CI standards, demonstrating both technical leadership and operational ownership. His strengths lie in lowering high-level ML frameworks to efficient backends, performance modeling for dataflow architectures, and bridging compiler frontends with silicon-specific codegen. Colleagues describe him as someone who moves seamlessly between mentoring, benchmark development, and the gritty details of devops and release automation.
3 years of coding experience