Preload GTM brain

Preload captures synchronized human demonstration data with wearable sensors for physical AI and robotics teams.

category

Physical AI training data

market profile

  • text: Robot manipulation teams need observations of contact and force that ordinary video cannot directly measure.
  • text: The company positions its collection, evaluation and training work across research labs, robotics developers, surgery and industry.

personas

  • title: Head of robot learning
    pains:
    • Human video demonstrations miss force and contact signals needed for dexterous manipulation.
    • Collecting synchronized real-world demonstrations at scale is operationally difficult.
  • title: Industrial automation lead
    pains:
    • Expert manual tasks are hard to translate into reliable robotic skills.
    • Promising demos must be evaluated against real task requirements before deployment.
  • title: Surgical robotics research lead
    pains:
    • Expert technique is difficult to measure objectively and capture for robotic research.

icp

  • text: Frontier AI labs and robotics companies training manipulation models with human demonstrations and access to robot evaluation hardware.
  • text: Industrial organizations with repetitive, contact-rich expert workflows and a budget for automation pilots.
  • text: Surgical robotics research groups exploring objective skill assessment and training data, subject to clinical validation.

message pillars

  • text: Measure the physical signal missing from video.
  • text: Collect synchronized demonstrations with multiple sensing modalities.
  • text: Connect captured human skills to practical automation.

objections

  • objection: Video datasets already scale cheaply.
    response: Pilot a contact-rich task and compare whether measured force and tactile data improve evaluation or learning enough to justify collection.
  • objection: Can the data transfer to our robot and tasks?
    response: Define the target embodiment and task, then assess demonstrations and real-hardware results jointly before expanding.
  • objection: How are surgical data and clinical outcomes validated?
    response: Request study design, permissions and evidence for the specific intended use; the site does not establish clinical clearance.

competitors

  • name: Nferent AI
    website: https://nferent.ai/
    why: Collects multimodal human skill datasets for robotics teams.
  • name: Wuji Technology
    why: Offers data collection gloves to robotics developers.
  • name: Proception
    why: Offers tactile glove hardware for robotics data collection.

keywords

  • robot training data collection
  • tactile glove robotics data
  • human demonstration datasets
  • physical AI multimodal data
  • robot manipulation force sensing
  • surgical skill capture

voice

  • Technical and research-led
  • Concrete about physical signals
  • Ambitious and deployment-oriented

signals

  • text: A robotics lab is training manipulation policies and needs more contact-rich human demonstrations.
  • text: An industrial team is exploring automation of expert manual workflows.

stack

  • egocentric video
  • tactile gloves
  • EMG

hiring

  • Software Engineer, Data Platform
  • Sensor Engineer (Wearables)
  • Robotics Scientist (Policy Learning)
  • Field Engineer, Surgical Data Collection

pricing

model: unknown

motion

sales_led

channels

  • LinkedIn
  • X
  • founder outreach

confidence

medium

Founders

  • name: Leo Li
    title: Founder
    role: Founder
    bio: Founder at Preload (YC F26). PhD student in Computational Neuroscience at Caltech. Previously Neuroengineer at Neuralink and software engineer at Amazon. BASc in Engineering Science (Robotics) from University of Toronto.
  • name: Surya Pandiaraju
    title: Founder
    role: Founder
    bio: Founder at Preload (YC F26). PhD candidate at Johns Hopkins BME, Waterloo BME grad Previously worked at - Sonera Magnetics - iota Biosciences - NeuroPace - Paradromics
  • name: Runming (Tony) Wang
    title: Founder
    role: Founder
    bio: Founder at Preload (YC F26). Biomedical Engineering PhD candidate at Johns Hopkins School of Medicine. Developed the first brain application of the Freeform Stimulator platform for focal cortical modulation. Performs animal neurosurgeries and fabricates neural recording and stimulation hardware. Previously conducted research at Emory/Georgia Tech on motor-unit coordination using EMG and motion tracking. BS in Neuroscience from Emory.

Back to YC companies