ABAB Pedia

Brandon McKinzie

Overview

Contents9 sections
Key facts

Topic: From physics to machine learning

McKinzie took part in an undergraduate research program at UC Davis in summer 2014. He studied physics and computing at UC BerkeleyUniversity of California, BerkeleyThe University of California, Berkeley (UC Berkeley) is a campus of the University of California system; the University of California was founded in 1868 through a merger of two institutions and moved to Berkeley, California, in 1873.Open the full entry, receiving the American Physical Society’s Helen Quinn Award in 2014 and the Berkeley physics department’s Laslett Scholarship in 2015. He wrote that taking Pieter AbbeelPieter AbbeelPieter Abbeel is an artificial intelligence and robotics researcher from Belgium, a professor at UC Berkeley and director of its Robot Learning Lab, known for research on robot apprenticeship learning and deep reinforcement learning. He co-founded Gradescope and Covariant and joined Amazon with part of the Covariant team in 2024.Open the full entry’s introductory AI class in early 2016 persuaded him to leave physics research. After all his first graduate applications were rejected, he spent months developing dialogue models, shared the code on GitHub and received a job offer from a natural-language-processing startup in Cambridge, MassachusettsMassachusettsMassachusetts (officially the Commonwealth of Massachusetts) is a state in the New England region of the northeastern United States; its capital and largest city is Boston. One of the original 13 states, it ratified the U.S. Constitution on February 6, 1788, the sixth state to do so. Its population in the 2020 census (April 1, 2020) was 7,029,917.Open the full entry.⁠[3][1][4]

Topic: Apple and OpenAI

After about a year he joined an NLP team at Apple, working on language modeling and training infrastructure. In early 2021 he moved to a team researching visual question answering and video captioning, contributed to on-device scene analysis for iOS and later joined the research organization led by Samy Bengio. In March 2024 Apple published the MM1 multimodal-model paper with McKinzie as first author. At OpenAI he helped build o3, which combined reasoning with tool use.⁠[4][5][2]

Interview: No Priors

Interview: No Priors

The October 31, 2025 highlights episode of No Priors replayed part of his discussion of reasoning models with Mitchell.⁠[6]

Advice Shared on No Priors

He suggested studying whether interaction with the world could help visual systems reduce uncertainty acquired from static training images. (25:09)⁠[2]

Related Organizations, Websites and Public Accounts

OpenAI: AI researcher. Official website: https://openai.com/ (opens in a new window).⁠[1]

Interview with No Priors

He gave an example of a visual blind spot: a model asked to read an analog clock often answered 10:10. He linked that tendency to the repeated time setting in online clock pictures.⁠[2]

Sources6 sources

  1. Brandon McKinzie (opens in a new window)linkedin.com↑
  2. No Priors Ep. 113 | With OpenAI's Eric Mitchell and Brandon McKinzie (opens in a new window)youtube.com↑
  3. Undergraduate Students (opens in a new window)calderon.faculty.ucdavis.edu↑
  4. About Me (opens in a new window)github.com↑
  5. MM1: Methods, Analysis & Insights from Multimodal LLM Pre-training (opens in a new window)machinelearning.apple.com↑
  6. No Priors Ep. 138 | The Best of 2025 (So Far) with Sarah Guo and Elad Gil (opens in a new window)youtube.com↑