Elias Frantar
Member Of Technical Staff at OpenAI
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolElias Frantar is a Member of Technical Staff at OpenAI and a researcher with a PhD from ISTAustria (advisor Dan Alistarh) who focuses on making massive ML models dramatically more efficient. He developed influential tooling and algorithms—most notably GPTQ and SparseGPT for low-bit quantization and sparsification—and led work on scaling laws for sparsely-connected foundation models during a DeepMind internship. His systems work includes Marlin, the first INT4xFP16 LLM inference kernel achieving near-ideal speedups at medium batch sizes, and QMoE, a framework that compresses trillion-parameter Mixture-of-Experts models to under 1 bit per parameter with custom CUDA kernels. Outside ML, he builds record-beating, high-speed Rubik’s Cube robots that have attracted millions of YouTube views, highlighting a knack for hardware-software co-design.
9 years of coding experience
2 years of employment as a software developer
Vienna University of TechnologyDoctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Institute of Science and Technology AustriaHigh School Diploma, Informatics, GPA: 4.00/4.00, High School Diploma, Informatics, GPA: 4.00/4.00 at Technologisches Gewerbemuseum
Github Skills (39)
pytorch10
llama10
inference10
tpu10
amd10
llm10
robot10
llmops10
cuda10
transformer10
gpt10
high-speed10
robotics9
xpu9
mlops9
Programming languages (2)
Github contributions (5)
efrantar/mirrcub3r
May 2018 - Feb 2020
Contributions:114 commits, 94 pushes, 1 branch in 1 year 8 months
IST-DASLab/spdy
Jul 2022 - Oct 2022
Code for ICML 2022 paper "SPDY: Accurate Pruning with Speedup Guarantees"
Contributions:1 review, 4 commits, 1 PR in 3 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.