Reinforcement Learning Engineer

The Robot Age; Austin, Northern (TX, KY); Hybrid; USD 120,000 - 180,000; Full time

Job summary

Apptronik is seeking an engineer to advance reinforcement learning for its Apollo humanoid platform in Austin. The role focuses on making learned policies work for whole-body loco-manipulation on physical hardware, spanning the loop from simulation to deployment.

The candidate will implement RL algorithms, transfer policies to real hardware, and collaborate with controls, hardware, and autonomy teams to improve training pipelines and robot behavior.

Qualifications

• Experience implementing reinforcement learning or robot-learning systems.
• Comfort with simulation-to-real transfer and physical robot constraints.
• Strong software engineering habits around experimentation and code review.
• Ability to collaborate across controls, hardware, and autonomy teams.

Responsibilities

• Implement reinforcement learning algorithms for locomotion and manipulation.
• Move policies from simulation into tests on physical humanoid hardware.
• Improve training pipelines for faster iteration and better transfer.
• Work with hardware, controls, and software teams on robot behavior.

Skills

RL systems

Sim-to-real transfer

Software engineering

Cross-team collaboration

Job description

What the role actually is

Apptronik is hiring a reinforcement learning engineer for its Apollo humanoid platform in Austin. The role is about making learned policies work for whole-body loco-manipulation on physical hardware.

The listing emphasizes the full loop from simulation prototypes to robot deployment, which makes this a real robot-learning role rather than a generic ML infrastructure position.

What you would work on

• Implement reinforcement learning algorithms for locomotion and manipulation
• Move policies from simulation into tests on physical humanoid hardware
• Improve training pipelines for faster iteration and better transfer
• Work with hardware, controls, and software teams on robot behavior

What they are asking for

• Experience implementing reinforcement learning or robot-learning systems
• Comfort with simulation-to-real transfer and physical robot constraints
• Strong software engineering habits around experimentation and code review
• Ability to collaborate across controls, hardware, and autonomy teams

Why this one is worth a look

Humanoid job postings can be vague, but this one names the behavior problem directly: dynamic locomotion and manipulation on hardware. That makes it a useful signal for candidates who want the hard part of humanoids, not just platform support.