Javier Sebastian Turek

Javier Sebastian Turek is an Argentine computer scientist working in applied AI, currently a Research Scientist at Earth Dynamics AI in the Portland, Oregon area. He represented Argentina at the 1999 International Olympiad in Informatics, winning a bronze medal (rank 81). He earned a B.Sc. (cum laude) and a Ph.D. in Computer Science from the Technion in Israel, where his dissertation on sparse representation modeling was advised by Michael Elad and Irad Yavneh. He spent roughly nine years as a Staff Research Scientist at Intel Labs in Hillsboro, Oregon, working on machine learning for neuroimaging, language modeling and perception, before leaving in early 2025. He has published at NeurIPS, ECCV, ACL and in signal-processing journals, and holds multiple Intel patents.

Notes8

• Left Intel in early 2025 after nine years, describing it as 'incredible'; joined an applied-AI startup in stealth that he framed as disrupting 'one of the oldest industries in the world'.Jan 2025
• Won a bronze medal representing Argentina at IOI 1999 (rank 81) - the origin of a competitive-programming path that led to a Technion CS PhD.Oct 16, 1999
• Authored the low-dimensional-structure paper (NeurIPS 2021) showing language-model representational geometry is reflected in brain responses - a bridge between his sparse-representation and fMRI/neuro-AI work.Jun 2021
• Google Scholar: 1426 citations, h-index 21, i10-index 38 - a solid research footprint concentrated in inverse problems, sparse representations and neuro-AI.2026
• PhD advisors Michael Elad and Irad Yavneh - Elad is a leading figure in sparse representation / dictionary learning, placing Turek in that school.2015
• Co-authored BrainIAK (Brain Imaging Analysis Kit) paper - contributor to an open-source neuroimaging analysis toolkit (github.com/brainiak).Jul 2021
• Featured in Intel's Hispanic Heritage blog, tying his Argentine roots to an immigration story he later told publicly with Ed Halla.Sep 2022