Jacob T
Graduate Engineer
Lund, Skåne County, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolJacob T is a Graduate Engineer with a decade of hands-on experience in software and machine learning systems, currently based in Lund and working at Arm. He progressed from a software engineering internship to a graduate role, bringing practical expertise in performance modeling for specialized accelerators. Jacob is an active contributor to the high-profile Apache TVM project, where he improved microNPU performance models, implemented conv2d op-specific costs, and expanded cascader matching logic. His MEng in Computer Science from the University of Leeds underpins a strong foundation in systems and compiler-aware ML optimization. Colleagues describe him as detail-oriented and pragmatic, focused on turning complex performance constraints into measurable improvements. He combines production engineering discipline with a research-minded approach to accelerator-aware compiler tooling.
10 years of coding experience
2 years of employment as a software developer
Master of Engineering - MEng, Computer Science, Master of Engineering - MEng, Computer Science at University of Leeds
Github Skills (13)
compiler10
tvm10
machine-learning10
deeplearning-ai10
performance-monitor10
tensor10
performance-analysis10
compiler-compiler10
deep-learning10
gpu10
python9
opencl8
roc7
Programming languages (1)
Github contributions (5)
apache/tvm
Dec 2021 - May 2022
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:
ML Engineer
ML EngineerContributions:19 reviews, 8 commits, 15 PRs in 4 months
Contributions summary:Jacob's contributions primarily involve the development and improvement of the performance model for the microNPU, a specialized accelerator within the TVM project. Their work includes adding performance modeling to the cascader, implementing op-specific performance code for conv2d operations, and addressing several bug fixes related to the performance model calculations. The user has also contributed to enhancing the cascader's capabilities by adding more Part matchers and options, demonstrating their focus on improving the efficiency and functionality of the NPU's performance estimation and optimization pipeline.
metalvulkancompilertensoropencl
jacobbohlin/tvm
Dec 2021 - May 2022
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Contributions:19 pushes, 10 branches in 5 months
cpugpu-programminggpu-accelerationtvmdeep-learning
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.