Arno Candel
Palo Alto, California, United States
Summary
🤩
RockstarArno Candel is a seasoned AI engineer and technology leader based in Palo Alto, currently a Member of Technical Staff at xAI with over a decade of experience building production-grade ML systems. Previously CTO and chief architect at H2O.ai, he led and heavily contributed to flagship platforms—including enterprise h2oGPTe, Driverless AI, and H2O-3—personally authoring tens of thousands of commits and helping h2oGPTe reach #1 on the GAIA benchmark. He pairs a PhD in computational physics (ETH Zurich, summa cum laude) and a history of building leadership-class HPC codes at SLAC with hands-on GenAI engineering, from integrating LLaMA and 4-bit quantization to enabling FlashAttention in h2ogpt. A Kaggle Master and pragmatic hacker, he is as comfortable designing distributed algorithms as shipping scalable, containerized agentic AI in production.
12 years of coding experience
Github Skills (21)
transformers10
spark10
integrate10
python10
machine-learning10
pyspark10
integrations10
tensorflow210
h2o10
tensorflow10
hugging-face10
gpt10
data-integration10
word2vec9
scala29
Programming languages (7)
Github contributions (5)
h2oai/h2ogpt
Apr 2023 - Jul 2024
Private chat with local GPT with document, images, video, etc. 100% private, Apache 2.0. Supports oLLaMa, Mixtral, llama.cpp, and more. Demo: https://gpt.h2o.ai/ https://gpt-docs.h2o.ai/
Role in this project:
Back-end Developer
Back-end DeveloperContributions:37 reviews, 54 PRs, 445 pushes in 1 year 3 months
Contributions summary:Arno primarily worked on the `generate.py` file, modifying the base model, default parameters, and generation output. They focused on incorporating and integrating various machine learning models, including LLaMA and other models into the generation pipeline. The contributions involved setting up and configuring the generation process, with emphasis on enabling 4-bit quantization and the use of Flash Attention. The user also implemented the saving of prompt/response data to JSON files.
aichatgptembeddingsgenerativegpt
h2oai/sparkling-water
May 2015 - Jun 2016
Sparkling Water provides H2O functionality inside Spark cluster
Role in this project:ML Engineer & Data Engineer
ML Engineer & Data EngineerContributions:50 commits, 15 pushes, 1 comment in 1 year 1 month
Contributions summary:Arno contributed significantly to the project by adding and refining code related to machine learning and data processing within the context of the Sparkling Water framework. They added a new example using Word2Vec for job title analysis. Additionally, the user updated and improved an existing Deep Learning example, and implemented a distributed TensorFlow deep learning MNIST demo for PySparkling. These contributions showcase a focus on integrating machine learning techniques with Spark and H2O.
develapiintegrationrsparklingh2o
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.