Member of Technical Staff, Machine Learning

Who are we?

Antioch is the simulation platform for robotics and autonomy. We allow teams to develop, evaluate, and generate synthetic data for their physical systems in high-fidelity cloud simulation, accelerated by 3D-native agents.

With Antioch, teams create precision-calibrated digital twins, generate testing environments, and run massively parallel simulation scenarios. By moving development out of the field and into software, Antioch drastically reduces dependence on expensive, slow, and incomplete real-world testing.

We work with leading robotics, autonomy, and perception teams to make the deployment of physical autonomous systems safer, faster, and more cost-effective.

Who are we looking for?

We're hiring a Member of Technical Staff, Machine Learning to be a core contributor to Antioch's ML team. You'll build the methods that make our simulations faithful to physical systems, and help drive the evolution from classical to learned simulation.

This is a hands-on technical role. You'll spend your time writing code, developing models, designing experiments, and shipping production capabilities alongside our simulation and core platform teams.

The central challenge is making simulation a reliable guide to real-world decisions. You'll combine physical models with learning from real data, demonstrate when simulated improvements transfer to hardware, and build a data feedback loop in which each deployment improves our models and calibration methods. You'll also contribute to Antioch's longer-term approach to learned simulation, including world models and diffusion models, grounded in measurable customer outcomes.

If you like the 0 → 1 grind and want your work to directly determine whether autonomous systems succeed in the real world, you'll fit right in.

What you'll do

• Measure sim-to-real and real-to-sim gaps across components and complete systems, with repeatable benchmarks, held-out validation, and clear estimates of uncertainty.

• Develop and ship calibration methods with our simulation team, combining physical models, system identification, and ML to improve fidelity against real-world data.

• Build self-serve workflows with our platform team to ingest diverse sensor, actuator, and system telemetry, calibrate digital twins, and validate results reproducibly.

• Establish a continuous data feedback loop: use deployment failures to guide data collection, improve models, and turn validated improvements into reusable platform capabilities.

• Advance Antioch's learned simulation roadmap, including world models and diffusion models, through practical experiments and production integration.

• Build and ship foundational ML models, pipelines, and tools as a core member of the team.

You'll work alongside Antioch's founders, simulation engineers developing physics and sensor models, platform engineers building cloud infrastructure and developer tools, and technical artists creating simulation worlds.

What we're looking for

You don't need every box checked. We're looking for strength in most of these:

• Deep applied ML experience, taking ambiguous problems from experiments to reliable production systems.

• Strong foundations in statistics, optimization, and experimental design, with sound judgment about uncertainty, data leakage, and generalization.

• Experience modeling physical systems in robotics, controls, system identification, sensor modeling, or a closely related field.

• Strong Python and software engineering skills for data pipelines, training, and evaluation, including effective use of coding agents and careful validation of their work.

• Experience turning messy, incomplete real-world telemetry into usable datasets, defensible models, and measurable improvements.

• Strong collaboration across disciplines, balancing scientific rigor with customer needs and time to ship.

Nice-to-haves

• Experience calibrating and validating robotic simulations using Isaac Sim/Lab, MuJoCo, or similar tools

• Experience with classical simulation, differentiable simulation, learned dynamics, world models, or diffusion models

• Experience with multimodal telemetry, signal processing, or sensing modalities such as cameras, radar, lidar, and infrared

• Experience shipping developer-facing ML workflows, operating GPU workloads, or building tools that agents can use reliably

You may be a great fit if…

• You're in NYC or open to relocating

• You're a high-agency builder: you don't wait to be told what to do, you go figure it out

• You care deeply about what you ship and reflexively take accountability

• You like the 0 → 1 environment and chose it on purpose

• You balance craft and pragmatism: you push for excellence, but ship what's needed now

• You're comfortable with uncertainty and love diving across the stack to solve a problem

• You're customer-obsessed and want the thing to work for real people

Why Antioch

• Real problems, real stakes. Our customers build physical autonomous systems that have to behave correctly in the real world, and our work directly affects whether they get there.

• High ownership. You'll own systems end-to-end rather than narrow tasks, with fast iteration and direct customer feedback.

• Team caliber. Work with a founding team who have previously built and sold companies, and worked at leading companies like Tesla Autopilot, Meta Reality Labs, and more.

• Build the company, not just the product. Early team members help define engineering culture, product direction, and how we win.

Interview Process

1. Application review

2. Introduction call

3. Technical Interview (a sample technical and programming interview that closely emulates work you'll do in the role)

4. Project Deep Dive (one of your past projects, in detail)

5. Onsite with the Antioch team in New York City