Title: The Imperative for Open Source Frontier AI - Oumi’s Platform & Strategy.

Speakers: Emmanouil Koukoumidis(CEO of Oumi PBC) and Oussama Elachqar(Member of Technical Staff at Oumi PBC)

Abstract: Open source is the best path to developing the most advanced, trusted, and versatile AI; the AI that would unlock industry applications and truly empower the sciences. Unfortunately though, frontier AI has become over the last 2+ years increasingly closed giving rise to significant concerns and risks. I will discuss the efforts that Oumi, alongside 20+ academics, recently started to put us on a better path by betting on and fostering open source frontier AI.  Lastly, I will make a call to action. To succeed, broad collaboration is needed across academia, the industry, and across the world. Greece could be a key player and leader in this.

Short Bios:

Emmanouil (Manos) Koukoumidis is the CEO of Oumi PBC focusing on democratizing foundation model research and development. He had previously led the science and engineering for Google Cloud’s Natural Language Services while also bootstrapping and leading the Google Cloud GenAI (PaLM/Gemini) efforts across a 300+ v-team covering model quality, Responsible AI, and all related infrastructure from inception all the way to General Availability. He has been working on GenAI systems since 2016 at Microsoft and then Meta. He holds a PhD from Princeton University where he set the early foundations for edge AI and a Diploma from the National Technical University of Athens. His research career includes positions at  MIT, Microsoft Research, SMART (Singapore) and CERN. His research has been featured in top-tier conferences, journals and news articles receiving a Best Paper Award and other distinctions with over 4,000 citations.

Oussama Elachqar is a Member of Technical Staff at Oumi PBC, where he’s focusing on democratizing foundation model research and development. He holds an M.Sc. from the Georgia Institute of Technology, and a Diplôme d'Ingénieur from INP-ENSEEIHT. Prior to Oumi, he led the vertical foundation model efforts for Healthcare at Apple, working all across the stack from pre-training, post-training, safety and performance optimizations. His expertise in generative AI systems spans nearly a decade, and his career includes leadership roles in Machine Learning at Microsoft, AI2 and Twitter, where he successfully delivered large-scale AI research projects to production. He specializes in scaling deep learning model training across HPC clusters and brings extensive experience in enterprise-scale AI deployment.

This talk will be held exclusively online through MS-TEAMS. More info below