Build the future of visual AI with us.
A 12-week fellowship for Master’s students and early-career researchers, embedded with our team on the world’s first Large Visual Memory Model.
Join the researchers teaching machines to see, remember, and understand.
A 12-week sprint at the
frontier of visual AI
The Memories.ai Research Fellowship is a 12-week intensive program for Master’s students and early-career researchers in computer vision, NLP, multimodal AI, and data systems. You won’t shadow anyone or work on side projects — you’ll be embedded with our research team on real, unsolved problems. We’re selecting 5–8 fellows for Fall 2026.
Duration
12 weeks · Sep–Nov 2026
Cohort size
5–8 fellows
Format
Remote-friendly, optional in-person
Final presentations
November 2026
Who we’re looking for
Currently pursuing a Master’s in computer science, AI/ML, or a related field — or a recent Bachelor’s graduate.
Hands-on experience with deep learning frameworks such as PyTorch or JAX.
Published, or working toward publications, in relevant venues.
Excited about video understanding, multimodal AI, or large-scale data systems.
You move fast, think deeply, and want to work on problems that haven’t been solved yet.
We value intellectual curiosity and a builder mentality over pedigree. If you’re doing interesting work — wherever you are — we want to hear from you.
Research at the frontier
Egocentric Video Understanding
First-person modeling: teaching AI to understand the world as humans experience it, from wearables and AR devices.
Large Visual Memory Models
Our flagship direction: architectures that give AI persistent, retrievable visual memory across long time horizons.
Real-Time Video Intelligence
Edge AI and streaming inference: processing live video feeds with low latency for real-world applications.
Multimodal Memory Systems
Cross-modal retrieval and reasoning: connecting what AI sees, hears, and reads into unified memory.
Data Infrastructure at Scale
Scalable pipelines to process, index, and retrieve massive video datasets — the backbone behind everything else.
Everything you need to do world-class work
Unlimited cloud compute
Full AWS and GCP credits. No rationing GPU hours — train the models you need to train.
State-of-the-art GPU clusters
Access to dedicated infrastructure for large-scale model training and experimentation.
Proprietary datasets
Work with Memories.ai video datasets and memory architectures you won’t find anywhere else.
1:1 mentorship
Weekly sessions with senior researchers, including former Meta Research Scientists.
World-class advisors
Access to industry experts and academic collaborators across our research network.