Eshed MargalitPronounced EH-shed Marg-uh-LEET, PhD

Computational Neuroscience and AI Researcher

I work on machine learning for cancer immunotherapy at Noetik. My background is in computational neuroscience, computer vision, and Neuro-AI. I also like creating tools for scientists and sharing educational resources for programming and statistics.

I'm broadly interested in discovering how complex systems (especially brains!) develop, and how to understand their function and dysfunction with machine learning.

You can find my academic CV or professional resume here, along with my blog.

What's new:

I gave a lecture on multimodal world modeling, and how we're using it at Noetik, at Stanford's CS25 Transformer's United V5 course. You can watch it on YouTube or view the slides.
May 2025
We (Noetik) released a technical report on our virtual cell work, Simulating Spatial Biology with Virtual Cells and Cellular Systems, alongside a companion app, Celleporter, that I developed.
Dec 2024
My core PhD research on brain-inspired neural network models of visual cortex is now published in Neuron. Read the paper or check out a recent talk I gave on this work.
Jul 2024
The first technical report describing our approach to world modeling at Noetik is out now: World Models as Simulators of Patient Biology.
Jul 2024
My brother and I launched milameter.run, an open-source Strava-integrated app that lets you sideload GPX files from a canine running companion and compare trajectories + statistics from your runs.
Jan 2023
I hit 200 consecutive weeks of detailed paper reviews on the free, open-source habit-tracker I built: Paper a Week. I no longer read papers weekly, but my notes on all 200 are publicly accessible on my profile.
Jan 2023
I defended my PhD thesis, titled A unified model of the structure and function of primate visual cortex.
May 2022
I organized a remote, mid-pandemic 6 week mini-course covering shell and Python basics for scientific research. The materials are available on the CNJCx website.
Sep 2020
My research on ultra-high spatial resolution fMRI of human visual cortex is out now in the The Journal of Neuroscience. Link to paper.
Apr 2020
I designed the websites for the Stanford Vision and Perception Neuroscience Lab (VPNL) and the Computational Neuroscience Journal Club, which I co-led at the time.
2020-2021
I made an interactive d-prime calculator for students in the undergrad psychology course I was a TA for. Try it out! Edit 2025: Claude helped me convert this to a NextJS app to simplify deployment.
Oct 2017