Ian Colbert
Senior Member Of Technical Staff at AMD
San Jose, California, United States
Summary
🤩
Rockstar🎓
Top SchoolIan Colbert is a Senior Member of Technical Staff at AMD with four years of industry experience focused on accelerating neural network inference at the intersection of hardware, software, and deep learning. He holds a PhD in Electrical Engineering from UC San Diego, where his thesis optimized end-to-end inference for image upsampling networks, and he has steadily progressed through engineering and research roles at AMD since 2018. His hands-on work includes FPGA-focused tooling—contributing to Xilinx's FINN project by improving hardware resource estimation and accumulator datatype logic—demonstrating deep competency in quantized NN inference and low-level optimization. Based in San Jose, he blends academic rigor with production engineering, often translating research ideas into practical compiler and hardware-aware software changes.
4 years of coding experience
10 years of employment as a software developer
University of California, San Diego
Github Skills (8)
fpga10
compiler10
compiler-compiler10
dataflow-programming10
dataflow10
neural-network9
quantization9
python8
Programming languages (2)
Github contributions (5)
Xilinx/finn
Dec 2022 - Jan 2023
Dataflow compiler for QNN inference on FPGAs
Role in this project:
ML Engineer
ML EngineerContributions:7 commits, 6 PRs in 1 month
Contributions summary:Ian's contributions primarily involve optimizing and refining the estimation logic for hardware resource utilization (LUTs) within the context of FPGA-based neural network inference. They updated the accumulator datatype calculations within matrix vector activation and vector vector activation operations. Their work also includes modifying the bit width calculations based on the accumulator datatype and checking for runtime writable weights. Additionally, the user updated the project's dependencies by changing the URL and commit information for qonnx.
dataflow-compilerfpgasinferencecompilerneural-network
i-colbert/brevitas
Jul 2022 - Mar 2025
Brevitas: quantization-aware training in PyTorch
Contributions:1 PR, 329 pushes, 82 branches in 2 years 8 months
pytorchdeep-learningquantization-aware-trainingmachine-learningtraining
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.