Team update

The OpenAI team is now 45 people. Together, we’re pushing the frontier of AI capabilities—whether by validating novel ideas, creating new software systems, or deploying machine learning on robots. We continue to look for creative, motivated researchers and engineers to help us achieve our goals.

Welcome to everyone who’s joined since our last team update!

• Aleks Kamko. Aleks(opens in a new window) is a recent graduate from Berkeley where he did research on distributed machine learning and taught Data Structures and Operating Systems for 2 years. Previously, Aleks was an intern at Stripe working on the Financial Operations team.

• Alex Ray. Alex(opens in a new window) previously built radio encryption systems for small satellites at Planet Labs(opens in a new window). Prior to that he worked on aerial robots at Airware(opens in a new window). He has a degree in Textile Engineering(opens in a new window) from NCSU, where he worked on a design for an inflatable habitat(opens in a new window) for living on Mars.

• Ankur Handa. Ankur obtained his PhD in real-time camera tracking at Imperial College London, then became a post-doc at the University of Cambridge to focus on scene understanding via his work on SceneNet(opens in a new window). He returned to Imperial as a Dyson Fellow to continue his research on scene understanding with SceneNet RGB-D(opens in a new window). His interests lie at the intersection of hardware, SLAM(opens in a new window), and deep learning.

• Bob McGrew. Bob formerly led engineering and product management at Palantir, having worked there for a decade as it grew from 10 to 1,500 employees. Before Palantir, Bob was a PhD candidate in the Stanford Artificial Intelligence Lab working on multiagent systems. At OpenAI, he focuses on robotics.

• Christopher Berner. Christopher(opens in a new window) was previously an engineer at Facebook, where he worked on Presto(opens in a new window), an open-source distributed SQL engine. He led the data warehouse deployment of Presto, and scaled it to hundreds of petabytes of data. In his spare time, he builds quadrotors and created Rospilot(opens in a new window), an autopilot companion system that provides real-time video streaming and other features, using commodity hardware.

• Erika Reinhardt. Erika(opens in a new window) was previously the Director of Product Engineering at Planet Labs(opens in a new window), where she spent the last four years building teams and software to support collecting and distributing high-frequency satellite imagery, ranging from mission control and manufacturing management, to image processing and platform design. She studied mechanical engineering and computer science at MIT. As OpenAI’s Engineering Director, Erika will help manage and grow the team.

• Jakub Pachocki. Jakub recently finished a PhD at CMU where he worked on the theory of optimization. He co-invented the fastest known algorithms for solving linear systems(opens in a new window) on graphs and computing the geometric median(opens in a new window). In the past, Jakub was a top competitive programmer, and has won the Google Code Jam(opens in a new window) and gained a gold medal in the ACM-ICPC(opens in a new window).

• Jeremy Schlatter. Jeremy(opens in a new window) is an engineer who previously worked at Mailgun and Google. Since joining OpenAI he has built the tight encoding(opens in a new window) implementation for OpenAI Universe’s go-vncdriver(opens in a new window) and tools for collecting human demonstrations of Universe environments.

• Jonathan Gray. Jonathan previously built a peer-to-peer file sync tool at AeroFS(opens in a new window), and was an early engineer at Magic(opens in a new window), an SMS-based virtual personal assistant. Prior to that, he studied Electrical & Computer Engineering at the University of Toronto. In high school and during his undergrad, he built software for cancer treatment at Sunnybrook Research Institute’s Focused Ultrasound Group(opens in a new window).

• Peter Welinder. Peter(opens in a new window) was an engineering lead at Dropbox, where he led the photos infrastructure, computer vision, and machine learning teams. Before that, he co-founded AI startup Anchovi Labs(opens in a new window) and did his graduate work at Caltech in computer vision(opens in a new window) and crowd sourcing(opens in a new window).

• Rachel Fong. Rachel most recently worked on data engineering at Locu(opens in a new window). In the past she’s designed robots, and built a haptic surgical simulator, a $50 desktop 3D scanner, and an NLP backend to extract data from uploaded documents to create personalized curriculums for learning language. She studied computer science at MIT.

• Shariq Hashme. Shariq founded an OCR startup, and previously worked at Gigster. He also builds DIY projects, like a $80 VR treadmill(opens in a new window), a portable shower, and a device to control a mouse cursor with a tongue. He studied Computer Science and Electrical Engineering at the University of Maryland, where he founded a hackathon club(opens in a new window) which won the first-ever Major League Hacking(opens in a new window) season.

• Szymon Sidor. Szymon(opens in a new window) previously worked on model-based reinforcement learning at Vicarious(opens in a new window), question answering at MetaMind, and on real-time file processing infrastructure at Dropbox. He earned his master’s degree from MIT, where he explored reinforcement learning approaches to multi-stage reasoning in natural language processing, and identified security flaws in Android(opens in a new window).

• Tom Brown. Tom(opens in a new window) was the cofounder and CTO of Grouper(opens in a new window), where he wrote algorithms that used Facebook data to match up millions of people so they could meet in the real world. Before that he studied computational cognitive science(opens in a new window) at MIT, and built early versions of the core server at MoPub(opens in a new window), now the world’s largest mobile ad server(opens in a new window).

• Yaroslav Bulatov. Yaroslav(opens in a new window) was previously an engineer at Google Brain and Google Streetview, where he implemented and trained large-scale neural networks. He designed the first system(opens in a new window) that outperformed humans at recognizing outdoor house numbers.

The following former OpenAI interns have joined us full-time: Catherine Olsson, Jonathan Ho, Paul Christiano, Peter Chen, Prafulla Dhariwal, Rein Houthooft, and Rocky Duan.