Preload

Multisensory systems (EMG+tactile+video) for training physical AI

Software Engineer, Data Platform

$150K - $225K•0.25% - 1.50%•San Francisco, CA
Job type
Full-time
Role
Engineering, Backend
Experience
2+ years
Visa
US citizenship/visa not required
Skills
Python, Distributed Systems
Connect directly with founders of the best YC-funded startups.
Apply to role ›
Leo Li
Leo Li
Founder

About the role

About Preload

Preload builds the data layer for physical AI. Robots learn mostly from video, but human interaction is much richer: touch, force, and muscle activity. We build a wearable multimodal stack (egocentric video, tactile gloves, EMG) to capture how people actually interact with the physical world, and turn it into datasets robotics teams can train on. We're a YC (F26) team of PhD founders from Johns Hopkins and Caltech with backgrounds at Neuralink, Paradromics, and other neurotech companies.

The role

You'll own the software end to end: from the apps that capture data on our wearables, to the pipelines that store, process, and serve large multimodal datasets.

What you'll do

  • Build data collection software that syncs video, tactile, and EMG streams from our devices
  • Build tools that field teams use to run collection sessions reliably
  • Design ingestion pipelines, storage, and metadata/labeling systems in the cloud
  • Build APIs and tooling to deliver datasets to ML teams and customers
  • Own reliability, monitoring, and scaling as collection grows

What we're looking for

  • 2+ years as a software engineer building backend or data systems
  • Strong Python; experience with cloud infrastructure (AWS/GCP) and data pipelines
  • Comfortable working close to hardware: device communication, streaming, timestamps
  • Experience with large video or sensor datasets is a plus
  • Experience with embedded, mobile, or robotics software is a plus

Why join

Early, in-person team in San Francisco. You'll be the first dedicated engineer on our data platform.

About Preload

Preload builds the data layer for physical AI.

Today's robots learn mostly from video, but people do much more than look. When we pick something up, tie a knot, or make a surgical incision, we use touch, force, and muscle activity, and video misses all of it. Preload is building a multimodal hardware stack that captures how people actually interact with the physical world. We collect that data at scale so robotics and physical AI teams can train models on it.

We're also exploring how this technology can advance surgery, where dexterity and precision matter most.

Who we are

We're a team of PhD founders from Johns Hopkins and Caltech, with backgrounds in brain-computer interfaces and neurotech at Neuralink, Paradromics, iota Biosciences, NeuroPace, and Amazon. We've built neural recording hardware, performed surgery, and studied human motor control. Now we're applying that experience to give machines human-level physical intelligence.

Why join us

  • Early and in person: we're an early-stage team in San Francisco, backed by Y Combinator (F26). You'll shape the product, the stack, and the culture from day one.
  • Hard, cross-disciplinary problems: the work spans hardware, sensors, signal processing, ML, and data infrastructure.
  • Real ownership: you'll ship hardware and software that go straight into the hands of people collecting data.

Who we're looking for

People who like building across hardware and software, move fast, and want to work on one of the biggest bottlenecks in robotics: high-quality physical interaction data.

Preload
Founded:2026
Batch:F26
Team Size:3
Status:
Active
Location:San Francisco
Founders