Nicholas Broad
Technical Account Manager at Together AI
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolNicholas Broad is a machine learning engineer and Technical Account Manager with eight years of experience building and deploying transformer-based models and production ML solutions from Stanford to San Francisco. He spent four years at Hugging Face accelerating client AI projects—custom LLMs, multi-hardware optimization (GPUs/TPUs/HPUs/IPUs), distillation and ONNX Runtime tuning—and now brings that hands-on model optimization experience to Together AI's developer-facing work. An active open-source contributor to the flagship huggingface/transformers repo, he has implemented robustness and memory-efficiency improvements and added features for Neftune and SDPA, signaling a focus on practical performance gains. His background blends deep research-grade fabrication and experimental work at Stanford with applied NLP and MLOps across pharma and startups, and he has taught enterprise transformer pretraining, demonstrating an ability to translate cutting-edge research into usable products. Notably, he pairs low-level engineering instincts (from semiconductor labs) with production-first ML engineering, making him effective at bridging research, tooling, and customer outcomes.
8 years of coding experience
8 years of employment as a software developer
Bachelor of Science (B.S.) Electrical Engineering, Bachelor of Science (B.S.) Electrical Engineering at Stanford University
English, Japanese
Stackoverflow
46reputation
1kreached
1answer
1question
Github Skills (12)
pytorch10
machine-learning10
transformer10
nlp10
deep-learning10
python10
natural-language-processing10
model-optimization10
pandas6
matlab6
flax4
jax4
Programming languages (11)
Github contributions (5)
huggingface/transformers
Jun 2021 - Sep 2022
🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
Role in this project:
ML Engineer
ML EngineerContributions:5 reviews, 5 commits, 13 PRs in 1 year 3 months
Contributions summary:Nicholas contributed to the Hugging Face Transformers library by implementing and improving various machine-learning related features. Their work includes modifying existing code to improve robustness and incorporating the use of variables instead of hardcoded strings. The user also made changes related to memory efficiency by switching to the chain() function instead of sum() for flattening lists. Furthermore, the user added functionality for Neftune and SDPA, suggesting a focus on optimizing model performance.
pythonbertspeech-recognitionstate-of-the-artflax
nbroad1881/need4speed
Feb 2023 - Feb 2023
Contributions:26 commits, 18 pushes, 1 branch in 9 days
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.