Miju Labs

← All buyers · Runway

J

Jimei Yang

Research Director · Runway

medium confidenceTier A · Talk to firstImage, video & audio modelsvideo · image

Yang is Runway's Research Director and one of the few people there with a long product history of shipping human-motion and character tools — at Adobe Research he built the body tracking in Character Animator and the motion capture behind After Effects' Project GoFigure. At Runway he works on video generative models and controllability. He is a technical requester: the person who would specify what a video-quality dataset actually has to contain.

Background

Yang spent years at Adobe Research working on human sensing, character animation and media synthesis, with shipped features in Adobe's creative tools rather than papers alone. His publication record runs past forty papers across CVPR, ICCV, ECCV, SIGGRAPH and NeurIPS, covering pose estimation, motion synthesis, image inpainting and 3D reconstruction.

He joined Runway and worked on the Gen-3 Alpha generation, now focused on video generation and control. He is based in Mountain View rather than Runway's New York headquarters. He speaks at the Runway AI Summit on the Research Q&A panel.

What they run now

Video generative models and controllability at Runway — the machinery that makes a generated shot do what a director asked. His Adobe background means he thinks in terms of shipped creative-tool features (rigging, motion transfer, tracking) rather than benchmark numbers, and his research area is precisely where automatic metrics fail and human judgement is required: motion plausibility, character consistency, physical coherence.

What they buy

Inference. Controllability research needs graded human judgement on whether a generated video obeyed an instruction — camera move, character action, continuity across shots — and on motion artifacts that no automatic metric catches. Likely specifies: instruction-adherence rubrics for video, motion-plausibility annotation, character-consistency pairs, and possibly motion-capture or reference footage. No vendor named publicly.

Career

  1. c. 2023 – presentResearch Director, Runway — Mountain View; Gen-3 Alpha and controllability work
  2. c. 2016 – c. 2023Research Scientist, Adobe Research — Character Animator body tracking; After Effects motion capture (Project GoFigure)
  3. c. 2011 – 2016PhD, computer science, University of California, Merced — Affiliation from publication record
  4. c. 2011–2016University of California, Merced, PhD, computer science (computer vision)

On the record

research aim · 2026

Describes his work as learning fundamental structures from large-scale visual data that connect real and virtual worlds. source

Interviews & talks

Why it matters here

High as a technical champion, moderate as a buyer. If Julian sells video-judgement data into Runway, Yang is the person who writes the spec and whose sign-off makes it real. His Adobe history also means he has been on the buying side of creative-tool data before and will recognise a serious rubric when he sees one.

How to reach

Technical, artefact-first. Speculation: send a worked rubric for instruction-adherence in generated video — the failure taxonomy, the disagreement rates between raters, what professional background predicts reliable scoring — using Gen-4.5 outputs as the material. He has a public X account and a personal site with a working email, and an academic's habit of replying to substantive mail. Do not open with company or price.

What we could not establish
  • Exact PhD institution and year (inferred from publication record).
  • Which Runway teams report to him and whether he controls any external spend.
  • Whether Runway's move to world models has changed his remit.

Why they are on the map

Listed as Research Director at Runway AI Summit 2026; senior owner of model research. source — evidence dated 2026

Needs the data for a model; writes the spec.

How Runway buys data

Gen-4.5 (Dec 2025) cited 'significant advances in post-training' and ranked via blind pairwise human preference on the Artificial Analysis arena. 2026 hiring includes 'Research Engineer, Data Foundations' (design multimodal datasets for world models), 'MTS, Data Infrastructure' (billions of rows of multimodal data), and a 'Sr/Staff PM, ML Research'. Past content licensing deals (Getty Images 2023, Lionsgate 2024) show a legal/licensing buying path. No named human-data vendor found.

Co-CEOs (Germanidis, Valenzuela) since Feb 2026, and a CTO (Sindi). Research is led by Research Director Jimei Yang with research leads under him; there are separate Data Foundations and Data Infrastructure groups, plus Runway Studios (in-house creatives) and a Chief Creative Officer.

Known vendors: lmarena (Artificial Analysis arena for public ranking; not a vendor)

Sources

  1. Runway AI Summit — San Francisco 2026 (speaker list)
  2. Ground Truth vol. 4 — Anastasis Germanidis (2026-09-18)
  3. Runway Research — Introducing General World Models
  4. Jimei Yang — personal site