Founding Robotics / ML Engineer

Founding Engineer, Robotics

• In short: funded, revenue-generating deep-tech AI company, still in stealth. This is a founding technical hire. You will own significant parts of the system and work directly with the founder. If you want scope and are prepared to earn it, read on. If you want a defined remit and a team around you, this is the wrong role.
• Location: Boston or Cambridge. Remote considered, but you need regular in-person time for hardware.
• Comp: $250,000 to $300,000 base, plus meaningful founding equity. Health, dental, vision, 401(k) with match.
• Experience: 5+ years hands-on, or a PhD with substantial real hardware work.
• Visas: open to sponsorship for the right person.

What the company does

• Spun out of a top research lab, backed by a specialist deep-tech investor. Patented technology, paying customers, and real revenue while still in stealth.
• Most AI today means bigger and bigger models trained offline. This team is doing the opposite: small systems that learn as they go, from live data, on a fraction of the usual compute. It works in the real, messy physical world. Customers are robotics companies, AI labs, and pharma.
• The founding team has shipped autonomous vehicles, personal robots, spacecraft, multimodal models, and safety-critical systems.

What you'd own

• Design and implementation of new learning architectures, from the maths through to running on a robot
• End-to-end ownership of technical problems: research direction, implementation, hardware integration, and whether it actually works in front of a customer
• Technical proposals, patent contributions, and customer evaluations
• Architecture decisions made directly with the founder

Requirements

You should be able to point to specific, named work for each of these:

• Physical hardware deployment at depth. You have taken learning or control systems onto real robots in unstructured environments and made them work reliably. Manipulators, mobile platforms, legged systems, vehicles, drones. Simulation-only experience does not meet this bar, and neither does having supervised someone else doing it.
• Founding-level ownership. You have carried technical work from ambiguous problem to shipped system without a spec, a manager, or an established team around you.
• Real depth in at least one of: reinforcement learning, control theory, robotic manipulation, or computer vision for robotics. Depth means you have implemented the methods yourself and understand why they fail.
• Python at a production standard, plus C or C++ for anything near the control loop.
• PyTorch or JAX at depth. Custom training loops, debugging models that train but do not work, reading and extending unfamiliar research code.
• Research signal. Publications at serious venues, patents, or open source that other people rely on. We want to see evidence of original technical work, not just delivery.

Strongly preferred

These are the areas the work actually sits in. Real experience in any of them moves you substantially up the list:

• Online, continual, or adaptive learning. Systems that keep updating after deployment rather than a frozen model. Lifelong learning, test-time adaptation, distribution shift on live data.
• Compute efficiency. Making models run inside hard budgets: quantisation, pruning, distillation, onboard inference, real-time constraints on limited hardware.
• Robotic manipulation. Contact-rich tasks, dexterous control, learned policies on real arms.
• Risk-aware or safe control. Uncertainty quantification, robust or safe RL, control under model error, failure that has a physical cost.
• Applied mathematics or signal processing. Sparse methods, sampling theory, information theory, approximation theory.

May not be best suited (unless you are keen on the above space), if:

• Your robotics experience is simulation only
• Your ML experience is fine-tuning or calling pretrained models
• You have been managing rather than building for the last two years

Thanks,

Team Amberes