Who is David Luan?

David Luan is a computer scientist, entrepreneur, and angel investor who is passionate about the intersection of machine learning and society. Luan also led the large model efforts at Google and has worked on AI-related projects at both Axon and OpenAI.

Category	Details
Full Name	David Luan
Nationality	American
Education	Bachelor of Science in Electrical Engineering and Computer Science from the Massachusetts Institute of Technology (MIT)
Early Career	Worked on machine learning and AI at various tech companies, Co-founded Dextro, a computer vision startup
Major Companies Founded	Dextro, Adept
Positions	Co-founder and CEO of Adept, Former Vice President of Engineering at OpenAI, Former Director of AI at Tesla
Business Milestones	– 2014: Co-founded Dextro, a startup focused on computer vision and video analysis using AI, which became a pioneer in real-time video content analysis. – 2016: Joined OpenAI as Vice President of Engineering, where he contributed to the development of large-scale AI models and research initiatives. – 2018: Became Director of AI at Tesla, overseeing the company’s efforts in developing AI for autonomous driving and other applications. – 2020: Left Tesla to start a new venture, Adept, aiming to build AI that can understand and automate complex processes and workflows. – 2021: Adept launched its initial products and services, focusing on enterprise AI solutions that enhance productivity and decision-making. – 2022: Adept secured significant funding from top-tier investors to accelerate its research and development efforts, expanding its team and capabilities. – 2023: Continued to lead Adept in advancing AI technologies, with a focus on making AI more accessible and practical for businesses across various industries.

Table of Contents

Education and early career

Luan was interested in computers from a very early age. Such was his passion that, between the ages of 8 and 13, he took night classes at Worcester State University in Massachusetts and earned a certificate in computer science.

As an adult, Luan earned a Bachelor of Science (Applied Mathematics and Political Science) from Yale University and was involved in various political and entrepreneurial affiliations.

Whilst a student, Luan also developed computer vision algorithms at iRobot Research and worked on machine learning for the Microsoft Academic Knowledge Graph (MAKG) project – a large RDF data set for scientific research that now houses over 238 million publications.

Dextro

Luan then founded Dextro in October 2011 and was also the company’s CEO. Dextro was a provider of an image and video recognition API that enabled clients to derive insights from objects appearing in their videos, photos, and streams. In fact, the API was the first to ship that could classify videos in real time.

Dextro was acquired by Axon in 2017 for $7.5 million, with Luan noting in a Medium post that the acquisition was “not only a natural fit culturally, but also technically. Our technology will serve as the foundation for Axon’s new suite of AI products and also bring state-of-the-art machine learning capabilities to an entire industry.”

Specifically, Luan’s tech would be used to process and analyze the immense amount of data generated by body-worn cameras in law enforcement. Luan continued at Axon for another 10 months and also established an AI Ethics Board with various institutions to better regulate facial recognition processes.

OpenAI

In December 2017, Luan joined OpenAI as Director of Engineering and then served as VP of Engineering where he worked with teams in generative modeling, algorithms, reinforcement learning, and language. Some of the models shipped during his time at the company include GPT-2, GPT-3, CLIP, and DALL-E.

He was also responsible for the company’s policy team with a focus on AI development, governance, and its long-term impacts on society. To that end, he set a broader research vision for OpenAI and was involved in miscellaneous tasks such as “promoting and coaching managers, technical work for the Microsoft fundraise, starting the recruiting team, rolling out our first perf system, etc.”

Google and Adept AI

Luan left OpenAI in September 2020 and joined Google as Director of Google Brain. He spent around 15 months at the company leading its large model efforts on projects that combined engineering and research.

Luan left Google in November of the following year as part of a wider exodus of top AI talent from tech companies who later started their own ventures.

Early in 2022, he subsequently co-founded Adept AI with fellow former Google employees Niki Parmar and Ashish Vaswani, as well as a cohort of employees who left their roles at both Google and DeepMind.

Luan started the company based on his belief in the power (and future) of transformers and their ability to perform a diverse set of general tasks: “The transformer was the first neural network that seemed to ‘just work’ for every major AI use case – it was the research that convinced me that general intelligence was possible”, he said in a statement.

Key takeaways:

David Luan is a computer scientist, entrepreneur, and angel investor who is passionate about the intersection of machine learning and society. Luan also led the large model efforts at Google and has worked on AI-related projects at both Axon and OpenAI.
Luan then founded Dextro in October 2011 and was also the company’s CEO. Dextro was a provider of an image and video recognition API that enabled clients to derive insights from objects appearing in videos, photos, and streams. When Dextro was taken over, Luan served with Axon for a short time before joining OpenAI.
Luan left OpenAI in September 2020 and then joined Google as Director of Google Brain. He spent around 15 months at the company leading its large model efforts before leaving to start Adept AI with Niki Parmar and Ashish Vaswani.