2022 Meta AI Program Mentors

October 6, 2022

Meta AI created the Artificial Intelligence (AI) Residency Program to give talented professionals hands-on experience with artificial intelligence research while working at a leading technology company. This one-year research training position is a great opportunity for those interested in honing their skills in areas such as image and video generation and AI-powered translation before applying to PhD programs in AI.

We asked current AI Residents to share their perspective on the program here. In this blog post, we introduce the Meta AI researchers serving as Mentors in the program.

Emily Dinan is a Research Engineer with Meta's FAIR lab in New York.

Emily Dinan is a Research Engineer in FAIR’s New York office and works on NLP research. “With our Resident, we are working on improving control over dialogue models such that they do not respond with toxic or offensive content. Meta is a great place to work on this type of research, as we get the opportunity to collaborate with experts in various fields across all of AI,” she says.

Emily loves working with the AI Residents — it’s really a unique opportunity to develop the skills of the next generation of AI talent. As a Mentor, Emily has learned a lot. “The Residents I’ve worked with have helped bring fresh perspectives to our research projects, and teaching AI skills has been a great way to challenge my own assumptions,” she says. When asked what advice she would give to future Residents, Emily suggests that future Residents make the most of their time by building connections with and learning from as many people as possible! There are so many opportunities to contribute to interesting projects and grow the skills that are most important to each person.

Akshara Rai is a Research Scientist on the Embodied AI team at FAIR’s Menlo Park office and works on AI for robotics. “My research is focused around building smart robots that can help us in day-to-day tasks. It is an extremely challenging problem, involving perception, language, control, and interaction, but also extremely rewarding to watch a robot understand and interact with the environment, and do something useful, like collect soft toys and put them in bags,” she says.

Akshara Rai's work focuses on the intersection of machine learning and control.

“Working at Meta and FAIR is a dream come true. Right now, we are still scratching the surface of intelligence in embodied agents, and Meta AI is a great place to work on such a far-looking project,” Akshara says.

“Both of my previous AI Residents started with backgrounds in AI and little knowledge of robotics. It has been highly rewarding to see them grow, enjoy, and understand the challenges of working with real robots,” she says. “Working with both has also given me fresh perspectives on learning for robotics, as well as AI research in general. My Residents are excited by the latest and greatest approaches, which results in a very interesting but grounded collaboration.”

Akshara advises Residents to figure out early on (with their Mentor) what they want out of the Residency, alongside their Mentor, and make a one-year plan based on that. It might help to make a one-year plan with intermediate goals and milestones to make sure that a project stays on track.

Ari Morcos is a Research Scientist with Meta's FAIR lab in Menlo Park, California.

Ari Morcos is a Research Scientist in FAIR’s Menlo Park office and works on understanding and evaluating deep networks. “Together with our Resident, we worked on two projects this year, one focused on whether we can train models to the same performance with less data and another focused on understanding the impact of different self-supervised learning approaches on the learned representations. Working on these projects at Meta is wonderful, because we are surrounded by world-renowned researchers in all these areas and have the freedom to pursue the ideas we think are most important,” says Ari.

Ari’s commitment to the program is palpable. “I believe that mentoring budding researchers who are new to the field is just as important as conducting good research. It’s been so rewarding to watch Residents grow their research skills and their confidence over the course of the program, and is an experience I look forward to each year. Working with Residents has taught me so much — how to be a better Mentor and tailor my mentorship style to each person’s unique needs, and how best to help Residents feel comfortable and confident in a fast-paced research environment. Since each Resident comes from a different background, Residents have encouraged me to look at projects from new and diverse angles.”

Adina Williams is a Research Scientist in FAIR Labs New York working on evaluating the performance of large language models. “There are many outstanding challenges for large language models, but one of the most pressing is the challenge of representing people from all groups fairly and safely. We have recently focused on better understanding the datasets and metrics used to measure social biases in large models, and together with our Residents, we have isolated a lot of noise and instability in these metrics and datasets,” she says.

“Understanding issues with current measurement techniques is the first step toward devising better ones, which can ultimately tell us how to improve our models in the future,” says Adina. “When working on projects with social import, it is important to have a large set of committed researchers with diverse backgrounds and viewpoints, as we are fostering at Meta, so we can better explore avenues for improvement quickly and inclusively.” Adina also finds mentoring really rewarding. “I’ve had to go back and re-earn (or learn for the first time) many technical details so that I could help my mentee. It’s been fun to (re)learn alongside a Resident!”

Adina Williams is a Research Scientist in FAIR Labs New York.

Adina’s advice to future Residents? “As a Resident at Meta AI, you are expected to be the main driver of your project, and it is up to the Resident to figure out what you need in order to have a successful project. For example, if you work better in groups, feel free to tell your Mentors—maybe they can find you a coding buddy! If you need a particular meeting cadence or more linear algebra review, let them know! Take some time to introspect on skills or experiences, and ask for them directly.”

Click here to learn more about the AI Residency and here to apply for the 2023 cohort.

Written By

Rashel Moritz

Program Manager