Linewise GTM brain

Linewise supplies real manufacturing design, machining and assembly data to frontier AI labs and robotics model developers.

category

Industrial data for physical AI

market profile

  • text: Physical AI teams need records of real industrial processes and human skills that are missing from public internet data.
  • text: Vendors compete on real-world capture, annotation and industrial coverage as robotics labs seek data beyond staged environments.

personas

  • title: Head of robotics research
    pains:
    • Insufficient variation in real factory demonstrations for generalist manipulation models
    • Hard to capture failed inspections and diverse operator techniques at production sites
  • title: AI training data acquisition lead
    pains:
    • Negotiating site access and participant consent for industrial footage
    • Evaluating provenance and dataset quality before a costly model training run
  • title: Engineering foundation model lead
    pains:
    • CAD outcomes alone omit design decisions, iterations and rejected designs
    • Connecting machining and inspection records to the original engineering intent

icp

  • text: Frontier AI labs and robotics companies training models on manipulation or industrial engineering tasks.
  • text: Teams able to specify and evaluate custom industrial datasets; company size and minimum contract are not public.

message pillars

  • text: Capture the entire journey from engineering to assembly.
  • text: Bring the variation of real factory work into model training.
  • text: Make each delivery traceable and verifiable.

objections

  • objection: Can we use the recordings with confidence in their rights and provenance?
    response: Review site agreements, participant consent and provenance documentation during the sample evaluation.
  • objection: Will the data meet our model's specifications?
    response: Agree the task, data format and quality criteria in the initial technical conversation and evaluate a sample.
  • objection: Why not use staged data or simulation?
    response: Compare representative production variation, including failures, against staged data on a held-out evaluation.

competitors

  • name: Scale AI
    why: Provides physical AI data collection and annotation to robotics model teams.
  • name: XDOF
    why: Collects robot training data using teleoperation and wearable sensors.
  • name: Claru
    why: Supplies captured and enriched robotics and physical AI datasets.

keywords

  • robotics training data
  • industrial physical AI datasets
  • manufacturing video data
  • robot manipulation data collection
  • engineering design training data
  • factory egocentric video

voice

  • Technical and concrete
  • Mission driven
  • Focused on traceability

signals

  • text: A robotics lab is building manipulation models and seeking varied real industrial demonstrations.
  • text: An AI team is evaluating engineering capabilities in CAD, CNC or manufacturing workflows.

stack

  • Egocentric video
  • UMI data
  • Hand pose annotation

hiring

  • CNC program lead
  • Mechanical engineering lead
  • Data delivery lead
  • VLM researcher
  • Agentic data infrastructure engineer

pricing

model: unknown

motion

sales_led

channels

  • LinkedIn
  • Research blog

confidence

high

Founders

  • name: Tanachart (James) Kujareevanich
    title: Founder/CEO
    role: CEO
    bio: James is the co-founder and CEO at Linewise. MBA from MIT. Former McKinsey Ops consultant.
  • name: Zhichu Ren
    title: Founder/CTO
    role: CTO
    bio: Zhichu is the co-founder and CTO at Linewise. MIT PhD. Built robotics automation for materials research.
  • name: Will Wenbo Zhang
    title: Founder/CPO
    role: CPO
    bio: Will is the co-founder and CPO at Linewise. MSE in AI & Robotics at UPenn GRASP Lab. Former founding engineer at Greyscale AI and Bestmow.

Back to YC companies