Lucas Atkins
CTO at Arcee AI | Leading open-weight foundation model development
About
Lucas Atkins serves as CTO and Head of Research at Arcee AI, where he spearheaded the proprietary training stack behind Trinity, Arcee's family of open-weight foundation models. With deep roots in machine learning that predate the current generative AI boom, he engineered next-generation text-to-speech systems for major automotive manufacturers including Volkswagen, BMW, Ford, and Hyundai.
Lucas has led the development of specialized language translation systems for the UAE and partnered with AMD in 2023 and 2024 to optimize enterprise GPUs for large-scale model training. He now channels this end-to-end experience into shaping Arcee's in-house models, focusing on efficient, open source and open-weight architectures.
His work aims to set the standard for American-made open intelligence, bringing together his expertise in production ML systems, hardware optimization, and foundation model training at scale.
What Lucas Talks About
Training 400B parameter mixture-of-experts models from scratch
Building proprietary training stacks for open-weight foundation models
Optimizing enterprise GPUs for large-scale model training
Developing efficient open source AI architectures
Enterprise text-to-speech and language translation systems
Talks2
Lucas Atkins at Data Council 2026
Trinity: Training a 400B MoE from Scratch Without Losing Your Mind
At a Glance
Is this you?
Claim this profile to edit your bio, add talks, and get verified.
Know a talk we're missing?
Get in Touch
Save for Later
Frequently Appeared With
Speakers who have appeared at the same events
Similar Speakers
Related by topic and shared stages
Explore related directories
Conferences, directories, and industries connected to Lucas Atkins