Job Description

What if your RL research didn’t just end in a paper, but helped autonomous aircraft make split-second decisions in places where GPS fails, comms go dark, and backup isn’t coming?

That’s the mission on the table, and it needs someone who’s ready to push reinforcement learning into the real world.

A deep tech startup is building autonomous VTOL systems for contested environments. And they’re hiring someone with serious research skills, someone who knows RL and wants to apply it where it actually matters.

Not to optimise ads. Not to fine-tune chatbot responses. But to guide high-stakes autonomous systems through uncertainty and dynamic environments.

The Role

You’ll be working on long-horizon tactical decision-making using RL and related methods, from reward design and policy training, to sim tuning and field testing.

You’ll collaborate with researchers and engineers on multi-agent problems, environment modelling, and real-world autonomy challenges. And yes, if publishing is important to you, you’ll be supported to keep doing it.

This isn’t a siloed postdoc. It’s a full-time, permanent role inside a company that values both deep thinking and functional systems.

Who This Might Suit

Maybe you’re just wrapping up a PhD in machine learning, robotics, or AI.

Maybe you’ve already done a postdoc but want to escape the 2-year treadmill.

Or maybe you’re an engineer with serious RL chops and no interest in the academic track, just the desire to build something that flies.

Either way, you should:

• Be confident reading RL papers and turning them into working code

• Understand how to shape reward functions and work with MDPs/POMDPs

• Care about how your work performs outside of simulation

• Be eligible to work in Australia (Citizenship or PR is a must)

• Be willing to work onsite (Melbourne or Adelaide)

Why It’s Different

• Full-time, permanent role

• Real systems, not just simulation or theory

• Room to grow - either into R&D leadership or deep technical IC

• High-agency, low-politics environment

• A team that values capability over credentials

Curious?

You don’t need to submit a polished CV to get started. Just reach out to Thaís on thais@theonset.com.au and we’ll take it from there.