Runtime.si GTM brain

Runtime.si builds simulation and evaluation infrastructure for robotics teams developing physical AI policies.

category

Robotics simulation and evaluation

market profile

  • text: Robotics teams need to test policies across conditions without depending entirely on scarce physical robot time.
  • text: The simulation-to-reality gap can make simulated policy performance a poor proxy for physical deployment.

personas

  • title: Robotics or embodied AI research lead
    pains:
    • Policy performance varies across tasks and environments
    • Physical test cycles are slow and difficult to reproduce
  • title: Simulation and evaluation engineer
    pains:
    • Building representative scenes and repeatable test harnesses
    • Finding edge cases before expensive real-world robot trials
  • title: VP of robotics engineering
    pains:
    • Need confidence in robot behavior before deploying into production
    • Hardware availability limits iteration throughput

icp

  • text: Robot developers and physical AI labs building manipulation policies, particularly in factory and logistics use cases; company size is not stated.
  • text: Teams already evaluating models or simulation environments and able to benchmark against real hardware are likely early design partners.

message pillars

  • text: Use simulation and search to scale robot intelligence.
  • text: Stress-test policies over diverse conditions.
  • text: Surface failures before deployment.

objections

  • objection: Will simulation scores predict performance on our robot?
    response: Run a matched simulation and physical evaluation and inspect the gap for the relevant tasks.
  • objection: Does it support our robot and simulator?
    response: Confirm embodiment, sensor, environment and integration support in an early access discussion.
  • objection: Why change our existing evaluation stack?
    response: Compare failure coverage and time to run a representative suite against the existing harness.

competitors

  • name: NVIDIA
    why: Isaac Sim provides robotics simulation and synthetic environments.
  • name: Runway
    website: https://runway.com/research/accelerating-robot-policy-evaluation
    why: Researches world-model-based robot policy evaluation.
  • name: One Robot
    why: Builds simulation environments for robot policy training and evaluation.

keywords

  • robot policy evaluation
  • robotics simulation platform
  • physical AI evaluation
  • sim to real robotics
  • robot manipulation benchmark
  • robot policy stress testing

voice

  • Technical and experimental
  • Direct and developer-oriented

signals

  • text: Team is hiring robotics infrastructure engineers or adding a policy evaluation function.
  • text: Company is moving a manipulation policy from laboratory testing toward a factory or logistics deployment.

stack

  • LIBERO benchmark
  • Nomad-Harness

hiring

  • Founding GTM
  • Founding Software Engineer, Robotics Infrastructure

pricing

model: unknown

motion

hybrid

channels

  • X
  • technical blog
  • open-source benchmarks
  • early access

confidence

medium

Founders

  • name: Chen Peng
    title: Co-Founder
    role: Founder
    bio: Co-founder, CEO at Runtime.si. Previously at Penn's GRASP and Penn Engineering working on model interpretability and brain foundation models. BA in Physics from UPenn.
  • name: Yim Ying
    title: Founder
    role: Founder
    bio: Co-founder at Runtime.si. Full-stack robotics engineer focused on robotics productization, manipulation, and manufacturing.MechE @ Penn.
  • name: Liam Zhong
    title: Founder/CTO
    role: CTO
    bio: Co-founder & CTO at Runtime.si. Graduate student / Researcher at UPenn GRASP Lab Program, CV, Control, and RL. Reviewer and author for CVPR, ECCV, NeurIPS, etc. Robotics@Penn.

Back to YC companies