Research - Grigori Fursin

Grigori Fursin

My R&D playground Mission LinkedIn Google scholar Twitter Medium GitHub EMail

Co-designing more efficient and cost-effective AI systems at FlexAI | modularizing and automating MLPerf @MLCommons | supporting open science and reproducibility @ACM & @IEEE & @HiPEAC | ex VP of MLOps @OctoAI | ex co-director of the Intel Exascale Lab | ex senior tenured scientist @INRIA | ex adjunct professor at the University of Paris-Scalay | PhD from the Unviersity of Edinburgh

My name is Grigori Fursin and I am an active open science advocate, reproducibility champion and open source contributor since 2007. I have an interdisciplinary background in computer systems, compilers, machine learning, physics and electronics. I also hold a PhD degree in computer engineering from the University of Edinburgh and I am a founder of cTuning.org (non-profit open science organization, 2014+), a founding member of MLCommons (2021+) and head of Cloud Services Labs at FlexAI (2024+).

My passion is to help researchers, engineers and students understand the SOTA AI, ML and Systems R&D and learn how to use it in the real world across rapidly evolving AI/ML models, data sets, software and hardware from different vendors - please see my ACM TechTalk and white paper to learn more about my vision. That's why I am glad to lead community developments of open-source tools, automation frameworks and platforms to fix the software/hardware mess, modularize complex AI systems, make them easier to use and automate their benchmarking, optimization and co-design to run AI, ML and other emerging workloads in the most efficient and cost-effective way in collaboration with MLCommons, ACM, IEEE and other organizations.

Please check a few recent presentations and publications if you want to learn more about my long-term projects, educational initiatives, open-source tools, Collective Knowledge Playground, Collective Mind workflow automation framework and portable, reusable and technology-agnostic automation recipes (CM4MLOps, CM4MLPerf and CM4ABTF) to support open science, reproducible research and artifact evaluation: ACM TechTalk'21, keynote at ACM REP'23, ArXiv white paper'24, overview in Philosophical Transactions of the Royal Society'21 and my reproducibility initiatives at ML and Systems conferences since 2014.

My current activities:

head of FlexAI Cloud Services Labs - please check flex.ai for more details.
organizer of reproducibility initiatives and artifact evaluation for AI, ML and Systems conferences and MLPerf benchmarks in collaboration with ACM, IEEE and MLCommons since 2013. I am leading the development of a common interface and automation language to make it easier to rerun and reuse code, data and experiments from published papers - see my ACM Tech Talk'21, ACM REP'23 keynote and white paper'24 for more details.
author of the Collective Mind workflow automation framework (CM) adopted by MLCommons and the Autonomous Vehicle Computing Consortium (AVCC) to modularize MLPerf benchmarks and make it easier to run them across diverse models, data sets, software and hardware from different vendors using portable, reusable and technology-agnostic automation recipes (CM4MLOps scripts). I donated this open-source technology to MLCommons to benefit everyone and continue developing it as a community effort. You can learn more about this project in this white paper. I thank our great contributors for their feedback and support.
founder and architect of the Collective Knowledge Playground - an educational initiative to learn how to co-design software and hardware to run AI, ML and other emerging workloads in the most efficient and cost-effective way across diverse models, datasets, software and hardware (trading off performance, power consumption, accuracy, cost and other characteristics). Please check this ArXiv white paper.

Brief summary of my past activities:

founder and co-chair of the MLCommons Task Force on Automation and Reproducibility to modularize and automate MLPerf benchmarks using my CM framework (white paper);
vice president of MLOps at OctoML where I prototyped the first version of CM and CM4MLOps together with the cTuning foundation before donating it to MLCommons to benefit everyone;
founder and chief architect of cKnowledge.io acquired by OctoML;
author of the Collective Knowledge technology (CK) powering cKnowledge.io;
author of the Artifact Evaluation and Reproducibility checklist (Unified Artifact Appendix) for ACM/IEEE conferences (see example of my artifact appendix at the end of this ASPLOS'24 paper "PyTorch 2: Faster Machine Learning Through Dynamic Python Bytecode Transformation and Graph Compilation");
co-founder of a CodeReef platform for universal MLOps;
founder in residence at Enterpreneur First;
co-director of the Intel Exascale Lab and tech.lead for performance analysis, optimization and co-design of high-performance and cost-effecitve computer systems;
senior tenured scientist at INRIA developing the foundations to co-design more efficient and cost-effective computer systems using auto-tuning, machine-learning and run-time adaptation;
research associate at the University of Edinburgh;
holder of the PhD in computer science from the University of Edinburgh with the Overseas Research Student Award (compilers, run-time systems and software/hardware co-design);
recipient of the European technology transfer award, ACM CGO test of time award and INRIA award of scientific excellence for my original research to use AI, ML, federated learning and collective tuning (cTuning) to automate development of high-performance and cost-effective computer systems and reduce R&D costs and time to market by an order of magnitude.

You can find some more details in my timeline.