Noah Verke
Co-Founder CTO at Ise AI
Los Angeles, California, United States
Summary
🤩
Rockstar🎓
Top SchoolNoah Verke is a Co-Founder and CTO with 11 years of engineering experience building production ML systems and developer tooling from Amazon to OctoML and now scaling Ise AI, a startup that generates editorial AI product photos to drive retail sell-through. He blends systems and ML expertise—optimizing compiler performance for Hexagon in the Apache TVM project and tuning LLM deployment configs in MLC-LLM—to squeeze latency and memory wins on specialized hardware. At OctoML he focused on ML systems and performance engineering, and earlier Amazon roles gave him large-scale software delivery experience. Based in Los Angeles, he pairs hands-on low-level optimization skills with startup product instincts, often surfacing practical benchmarks and tests to prove performance improvements.
11 years of coding experience
7 years of employment as a software developer
Bachelor’s Degree Computer Engineering, Bachelor’s Degree Computer Engineering at Loyola Marymount University
English
Github Skills (23)
tvm10
c-language10
language-model10
system-configuration10
compilation10
python10
llm10
custom-configuration10
hexagons10
compiler-compiler10
performance-optimization10
compile10
compiler10
configurations10
yml-configuration10
Programming languages (1)
Github contributions (5)
apache/tvm
Jun 2022 - Jan 2023
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:
Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:63 reviews, 19 commits, 30 PRs in 7 months
Contributions summary:Noah's contributions primarily focus on optimizing the TVM compiler stack for the Hexagon architecture. They made several changes to improve the performance of HVX workloads, including updating concurrency settings and adding VTCM loading capabilities. Additionally, they introduced new tests and examples to showcase and benchmark the effectiveness of async DMA pipelining and data transfer strategies for Hexagon. This work directly involves improving the efficiency of machine learning workloads on specialized hardware.
metalvulkancompilertensoropencl
mlc-ai/mlc-llm
Aug 2023 - Oct 2023
Universal LLM Deployment Engine with ML Compilation
Role in this project:ML Engineer
ML EngineerContributions:6 reviews, 5 PRs, 12 comments in 1 month
Contributions summary:Noah primarily contributed to the model and configuration aspects of the MLC-LLM project. They made changes to the Llama2 model, including caching and using `max_sequence_length`, indicating an understanding of model architecture and optimization for sequence length. The user also modified the core and relax_model files, incorporating changes related to configuration overrides, demonstrating a focus on model build and configuration. These contributions suggest a role in model deployment and configuration.
language-modelllmmachine-learning-compilationtvm
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.