Miju Labs

← All buyers · Apple Foundation Models (AFM)

Titus Barik

Researcher, Apple (HCI and software engineering) · Apple Foundation Models (AFM)

medium confidenceTier A · Talk to firstSecond-wave labsUI

The senior author on Apple's designer-feedback UI-generation paper and, by the look of his publication record, the person who anchors Apple's Seattle-area HCI group where designer-in-the-loop model work happens. He is also a CHI 2026 subcommittee chair and an ICSE 2026 programme committee member, which makes him a well-connected node in the design-and-code research community rather than a procurement contact.

Background

Describes himself simply as a researcher at Apple working at the intersection of HCI and software engineering. His visible output is conference-shaped: the CHI 2026 designer-feedback paper with Jason Wu, Amanda Swearngin, Arun Krishna Vajjala, Alan Leung and Jeffrey Nichols, and a second CHI 2026 paper with Priyan Vaithilingam, Alan Leung and Nichols on instantiating UI components with distinguishing variations. He carries service load in both communities — subcommittee chair at CHI 2026, research-track PC at ICSE 2026 — which is the profile of a mid-to-senior industry researcher with a strong academic network.

What they run now

Design-and-code research inside Apple: how models can be taught to generate better UIs, and how UI components should be presented and varied. In the designer-feedback work his group collected roughly 1,500 annotations from 21 designers using comments, sketches and direct manipulation, captured the rationale behind each critique, fine-tuned several LLMs on it, and showed human judges preferred the result over ranking-trained baselines and GPT-5. That is the closest thing in the literature to a specification for a design-critique dataset product.

What they buy

Inference. Apple research groups typically run design studies with internal designers and vendor-supplied participants; the designer-feedback study used 21 designers, which is participant-panel scale rather than data-contract scale. If his group scales that method — and the paper's results give it every reason to — the purchase becomes: recruited professional designers, a critique-capture protocol (comment, sketch, direct edit), rationale text, and quality control over it. Apple's foundation-model reports separately describe collecting large text-image preference datasets for reward modelling with 'internal and external' human evaluation, vendors unnamed.

Career

  1. presentResearcher (HCI / software engineering), Apple — Senior author on the CHI 2026 designer-feedback paper.

On the record

Designer-aligned training · 2026-02

Co-author of the claim that designer-native critique produces better UI generators than ranking feedback. source

Interviews & talks

Why it matters here

He is the most senior public owner of the designer-feedback line of work still at Apple, and therefore the person most likely to commission a scaled version of exactly the dataset Julian wants to sell. The caveats are real: Apple is notoriously hard to sell into from outside, research participants are usually procured through existing approved vendors, and none of Apple's human-data suppliers are ever named.

How to reach

Academic channel, not sales channel. Speculation: the credible move is to build on his paper publicly — for example a reproduction or extension with a larger, external designer panel and published inter-rater agreement — and to meet him at CHI, where he holds a chair role. An opener that works is a methodological question about where the 21-designer protocol breaks at scale (rater drift, rationale quality, sketch capture), with an offer to run a pilot. Do not open with vendor capabilities or pricing.

What we could not establish
  • His career history before Apple and his education — barik.net blocks automated fetching and I found no reliable secondary source.
  • His exact internal title and seniority at Apple.
  • Whether Apple has scaled the designer-feedback collection beyond internal designers, and through which vendor.

Why they are on the map

Senior author on designer-feedback UI-generation paper (CHI 2026) source — evidence dated 2026-02

Shapes what 'good data' means.

How Apple Foundation Models (AFM) buys data

AFM 2025 report (Jul 2025): collected large text-image preference data to train a reward model; human graders disagree on 20-30% of pairs; 'internal and external' human eval, vendors unnamed. AFM v3 (Jun 2026) uses in-house human graders. Apple researchers (CHI 2026) collected 1,500 designer annotations from 21 internal designers to fine-tune UI-generation models - direct design-feedback data work.

Dec 2025: Giannandrea retired; Amar Subramanya (VP AI) under Craig Federighi runs AFM, ML research and safety; Zhifeng Chen leads the foundation-models team. Human grading appears in-house; HCI/UI research group in Seattle (Nichols, Barik, Swearngin).

Sources

  1. ICSE 2026 profile — Titus Barik, Apple, HCI and software engineering
  2. Apple ML Research — Apple at CHI 2026 (papers and roles)
  3. Improving User Interface Generation Models from Designer Feedback (arXiv 2509.16779)