Developer Advocate Engineer

About the role

Dexmate is building the foundation for physical AI - combining a new generation of robots with a universal Physical AI OS, making robots as easy to build and deploy as software. Today, robotics is fragmented, slow, and closed: most builders are forced to reinvent the same stack again and again, and most ideas never make it past the prototype stage. We exist to change that. Our mission is to democratize robotics by lowering the barrier to entry, delivering a plug-and-play platform for developers, researchers, and enterprises, and cultivating an open ecosystem that accelerates the evolution of physical AI. If you want to help shape the next layer of human capability - and believe the future of robotics should be built together, not in isolation - we'd love to build it with you.

The role

Many developers have never deployed a model that moves something in the physical world. The gap between "it works in training" and "it works on a robot" is enormous.

That's your job.

You'll be the person who helps AI/ML developers understand what happens when their models leave the data center and run on a humanoid robot: latency constraints, sensor noise, sim-to-real transfer, on-device inference, closed-loop control, and more. You'll build the sample projects, write the tutorials, and create the content that makes Dexmate the platform serious AI engineers choose when they want to work on physical AI.

This is an engineering role first. You write code every week. The talks and tutorials come from building, not the other way around.

What you'll do

• Build and publish sample projects that show AI/ML engineers how to train, fine-tune, and deploy models on the Dexmate platform, demonstrating end-to-end physical AI workflows.
• Write and publish technical tutorials weekly - step-by-step guides, architecture explainers, and deployment walkthroughs written for engineers who know ML but are new to physical AI.
• Own the SDK documentation for AI/ML workflows: quickstart guides, API reference, Python SDK samples, kept current within 48 hours of any platform change to ensure accuracy and usability.
• Answer developer questions daily in Discord and GitHub Discussions - no question unanswered within 24 hours, fostering a responsive and supportive community.
• Build reference integrations with foundation AI model providers and publish architecture guides for running their models on Dexmate robots, highlighting performance and tradeoffs.
• Speak at AI/ML conferences 3-4 times per year - NeurIPS, ICLR, ICML, CoRL, and similar - to share insights and showcase platform capabilities to broad technical audiences.
• Run live demos for developers, partners, and enterprise prospects, translating complex physical AI concepts into clear, engaging experiences.
• Surface model integration friction and missing platform capabilities to engineering weekly, creating a tight feedback loop that improves the developer experience and platform roadmap.

Who you are

• You write Python fluently and have real ML engineering experience - model training, fine-tuning, inference optimization, or ML infrastructure. You've shipped models that ran in production, proving your ability to deliver reliable, scalable solutions.
• Curious about the physical world. You don't need a robotics background, but you find the question "what happens when this model controls a robot arm" genuinely interesting, not intimidating, and you approach it with a learner's mindset.
• You've published technical content that got traction - a GitHub repo people starred, a tutorial people bookmarked, a blog post that circulated in ML communities - demonstrating your ability to create impactful, widely appreciated resources.
• You write code other engineers want to copy. Clean, documented, opinionated about the right way to do things, ensuring that your work serves as both a practical guide and a source of inspiration.
• You can write a clear getting-started guide for a developer who just signed up and deliver a credible technical talk to a room of ML researchers, adjusting depth and register for the audience without losing substance.
• 3+ years of ML engineering experience - model development, training infrastructure, inference, or MLOps - providing a solid foundation in the practices and tools that underpin modern AI systems.
• You publish on a schedule. The failure mode this role avoids is someone who plans great content but never ships it; you prioritize execution and consistent delivery of high-quality materials.
• Experience with vision-language-action models or embodied AI research, giving you deeper insight into the challenges and opportunities of physical AI systems.
• Hands-on sim-to-real transfer work, allowing you to bridge the gap between simulation environments and real-world robot behavior effectively.
• Familiarity with Isaac Sim, MuJoCo, Drake, or Genesis, equipping you to work with leading simulation platforms used in robotics research.
• An existing technical blog, GitHub, or YouTube presence with real traction, demonstrating your ability to build an audience and share knowledge.
• A record of open-source ML contributions, showing your engagement with the broader community and commitment to collaborative progress.
• Experience deploying models on edge or embedded hardware, providing practical insight into the constraints and opportunities of real-device execution.

If you've ever wanted to be the person who explains physical AI to the world - this is that job.

Practical notes

Location: Fremont

Engagement: Full-time

There are no specific hours, travel requirements, visa conditions, or application deadlines mentioned in the source material. The focus is on the responsibilities, expectations, and qualifications for the role as defined by the available content.