Oleksandr Maksymets
AI Researcher At Multimodal LLMs Pre-training In Llama, Meta Superintelligence Labs
Menlo Park, California, United States
Summary
🤩
Rockstar🎓
Top SchoolOleksandr Maksymets is a multidisciplinary AI researcher and tech lead with 12 years of experience building large-scale multimodal systems, currently architecting next-generation vision-language encoders for Llama at Meta Superintelligence Labs. He has driven state-of-the-art ImageNet and retrieval results while cutting multibillion-sample training runs from months to weeks through data-pipeline, memory-layout, and parallelization optimizations that yield 30–50% speedups. Prior work at FAIR includes leading Visual Cortex and Habitat efforts that boosted embodied navigation benchmarks and enabled sim-to-real robotics evaluation, and he contributed core DevOps and CI infrastructure to the widely used Habitat simulator. A PhD-trained thinker who moves comfortably between research and production, he routinely runs thousand-GPU experiments, integrates privacy-compliant petabyte-scale data, and aligns cross-org roadmaps to eliminate duplication. Notably, his combination of bench-level SOTA results and hands-on systems engineering—automating GPU testbeds and conda builds—makes him rare among ML researchers who also ship robust, scalable infrastructure.
12 years of coding experience
10 years of employment as a software developer
Mathematics and Computer Science, Mathematics and Computer Science at Chernivtsi Lyceum #1Doctor of Philosophy (Ph.D.) Computer Science, Doctor of Philosophy (Ph.D.) Computer Science at Taras Shevchenko National University of Kyiv
English, Ukrainian
Github Skills (24)
pytest10
python10
testing10
configuration-management10
conda10
reinforcement-learning10
cicd10
automation10
deep-learning10
automation-testing10
computer-vision10
automations10
simulator9
c-language9
dockers9
Programming languages (8)
Github contributions (5)
facebookresearch/habitat-lab
Apr 2019 - Dec 2022
A modular high-level library to train embodied AI agents across a variety of tasks and environments.
Role in this project:
Backend Developer
Backend DeveloperContributions:13 releases, 256 reviews, 98 commits in 3 years 8 months
Contributions summary:Oleksandr primarily contributed to the development of the baseline agents, incorporating improvements such as agents suitable for challenge submissions. They modified core files related to agents, config improvements, and dataset management. Furthermore, the user refactored code related to the config system, which improves parameter sweeping and dataset specification.
pythonembodied-aideep-reinforcement-learningsimulatorai-agents
facebookresearch/habitat-sim
Jun 2019 - Jul 2021
A flexible, high-performance 3D simulator for Embodied AI research.
Role in this project:DevOps Engineer & Test Automation Engineer
DevOps Engineer & Test Automation EngineerContributions:5 releases, 48 reviews, 38 commits in 2 years 2 months
Contributions summary:Oleksandr contributed significantly to the repository's infrastructure and testing capabilities. Their work included setting up a continuous integration (CI) environment using GPU machines for testing, integrating speed regression tests, and adding a script to collect environment information for debugging purposes. Furthermore, they contributed to the automated documentation building and website content updates. The user also implemented conda builds for different configurations and operating systems, along with the addition of automated cleanup scripts for conda package management.
roboticsembodied-aicomputer-visionsimulation3d-simulator
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.