Amazon opens new AI lab in San Francisco with focus on long-term research bets

From left to right: David Luan, VP of Autonomy and Head of Amazon’s AGI SF Lab, and Pieter Abbeel, Amazon Scholar, Robotics.

Today, we are excited to announce the formation of the Amazon AGI SF Lab, a new dedicated team based in San Francisco. The initial focus of our lab will be to develop new fundamental capabilities to enable useful AI agents that can take actions in the digital and physical world. In other words, we enable practical artificial intelligence that can actually do things for us and make our customers more productive, empowered and satisfied. Our work will build on Amazon’s broader AGI team, which recently introduced Amazon Nova, a new generation of advanced foundation models (FMs).

Introducing Amazon Nova, a new generation of state-of-the-art foundation models that deliver boundary-breaking intelligence and industry-leading price performance. Amazon Nova Pro is a highly suitable multimodal model with the best combination of accuracy, speed and cost for a wide range of tasks. Amazon Nova Lite is an inexpensive multimodal model that is lightning fast for processing images, video, documents and text

The Amazon AGI SF Lab is designed to empower AI researchers and engineers to make major breakthroughs with speed and focus toward this goal. Our philosophy combines the agility of a startup with the resources of Amazon. By keeping the team lean, we are able to maximize the amount of calculations per person. Each team in the lab has the autonomy to move quickly and the long-term commitment to pursue high-risk, high-payoff research.

The new lab is seeded by our team recently hired from Adept and will leverage their pioneering work in building AI agents that can handle complex workflows using the same tools we use as humans, such as computers, web browsers and code interpreters. Our initial focus is on several key research bets that will enable AI agents to perform real-world actions, learn from human feedback, self-course correct, and derive our goals. In particular, we are really excited about work combining large language models (LLMs) with reinforcement learning (RL) to solve reasoning and planning, learned world models, and generalization tools for physical environments.

Fast: "A cave lit by shafts of light revealing hidden subterranean pools, the camera rolls counter-clockwise." Made using Amazon Nova Reel.

Amazon Nova Canvas and Amazon Nova Reel use diffusion transformers to deliver studio-quality visual content.

We are looking for a few dozen passionate, talented people to join our team – not only AI experts who have trained state-of-the-art models, but also graduates from other disciplines who will bring fresh thinking to the field , such as physics, mathematics or quantitative economics, regardless of experience level. If you are interested in our particular philosophy of AI advancement, please contact us at AGI-SFLab-Jobs@amazon.com.

The two of us have seen many eras of AI development. We are entering an exciting new era where AI agents are the next playing field; the right research bets can reinvent what is possible with AI. We’d love for you to join our new lab and build it from the ground up!

Leave a Comment