Arnav Garg
Technical Lead Manager - AI at Rubrik
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolArnav Garg is a Technical Lead Manager (AI) based in San Francisco who builds agentic, customer-facing AI products and leads small ML teams to deliver 0→1 systems at Rubrik. He previously led Predibase’s LLM efforts—co-creating Turbo LoRA for faster inference, shipping reinforcement fine-tuning offerings, and authoring practical resources like LoRA Land—bringing a mix of research-informed engineering and production pragmatism. An active open-source maintainer of Ludwig, he contributes backend and ML improvements that make low-code multimodal training more robust and user-friendly. Earlier roles at Atlassian and startups honed his product-focused ML skills across recommendation, ranking, and growth models, and he holds a BS in Computer Science from UCLA. He’s particularly interested in tooling and workflows that let teams iterate on LLMs at scale, and has a knack for squeezing large-model performance gains out of pragmatic systems work.
3 years of coding experience
7 years of employment as a software developer
The International School BangaloreUniversity of California, Los Angeles
Chinese, Spanish, Sanskrit, Hindi, English
Stackoverflow
286reputation
220kreached
16answers
1question
Github Skills (15)
hyperparameter-optimization10
pytorch10
machine-learning10
python10
back-end-development10
ml10
numpy9
llm9
data-science8
neural-network6
keras6
loss-functions6
tensorflow6
gpu6
theano6
Programming languages (3)
Github contributions (5)
ludwig-ai/ludwig
Jun 2022 - Jan 2023
Low-code framework for building custom LLMs, neural networks, and other AI models
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:3 releases, 1040 reviews, 288 commits in 7 months
Contributions summary:Arnav made several contributions focused on improving the Ludwig AI library, particularly concerning the processing of numerical and text-based features. Specifically, the user addressed warnings related to NumPy in the trainer and incorporated packing versions for library compatibility. In addition, the user expanded the hyperopt functionality by adding shared parameter capabilities and also fixed postprocessing calculations within the model evaluation. These contributions improved the library's performance and addressed various data type concerns in model inference.
fairness-mlpythonframework-learningdeep-learning-frameworknatural-language-processing
arnavgarg1/arnavgarg1
Aug 2022 - Feb 2025
Contributions:22 pushes, 1 branch in 2 years 6 months
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.