Suggestions
Dmytro Dzhulgakov
CTO and Co-Founder at Fireworks AI
Dmytro Dzhulgakov is a prominent figure in the field of artificial intelligence and machine learning, currently serving as the CTO and Co-Founder of Fireworks AI since October 2022.2 His career has been marked by significant contributions to AI infrastructure and deep learning frameworks.
Career Highlights
Meta (Facebook): Prior to his role at Fireworks AI, Dzhulgakov held several key positions at Meta (formerly Facebook) from 2011 to 2022:
- Tech Executive (2018-2022): Led the PyTorch and Meta AI Platform, overseeing a team of 400+ engineers.2
- Senior Staff Software Engineer (2016-2018): Co-founded Facebook's first production AI Platform and was a core developer of the Caffe2 deep learning framework.2
- Staff Software Engineer (2014-2016): Founding member of the Ads Personalization team, developing large-scale machine learning systems.2
Notable Achievements:
- Co-creator of ONNX, an initiative to improve AI development interoperability.1
- Core maintainer of PyTorch, leading its development through version 1.0 and beyond.2
- Instrumental in consolidating AI toolchains across research and product teams at Meta, significantly reducing deployment time for AI advancements.2
Education and Early Career
- Master's and Bachelor's degrees in Applied Mathematics from National Technical University "Kharkiv Polytechnic Institute" in Ukraine.2
- Co-founded QBit, a nonprofit organization promoting programming among students in Ukraine.2
- Worked as a Software Developer at NPF Informatica, developing SCADA systems for industrial process control.2
Expertise and Recognition
Dzhulgakov is known for his expertise in:
- Deep learning frameworks and AI infrastructure
- Large-scale machine learning systems
- AI platform architecture and scalability
He has been recognized for his skills in competitive programming, having been ranked in the Top 20 on TopCoder.12
Dmytro Dzhulgakov continues to be active in the AI community, speaking at conferences and sharing insights on efficient LLM serving and AI scaling.34