AI Tech Lead

About Georgian

Georgian is a Toronto-based venture capital firm that partners with high-growth technology companies across cybersecurity, AI software, robotics, defense. We manage about $6B in assets and we have invested in more than 80 companies that we support with capital and technical capability.

Our in-house AI lab works directly with portfolio companies to provide a differentiated experience across the investment lifecycle, and drives the adoption of AI for our internal operations.

AI Tech Lead - Physical AI & Robotics

As a senior technical lead in Georgian’s AI Lab, you’ll work on the brain of the robot: the models it thinks with, the data it learns from, the simulators it practices in and the systems that get all of that onto real machines. The job has three parts: research what’s coming next, work inside our portfolio companies to get robot-learning systems from pilot into production and help our investment team judge physical AI companies. This is a software job. You won’t design actuators or own the hardware, but you’ll know enough about the body to understand what it will and won’t let the brain do.

Why this role exists

Physical AI is a core investment thesis area for us. Our portfolio includes Physical Intelligence, Generalist, Bedrock Robotics, Path Robotics and Vention. We think most of what separates these companies is the brain, not the body: how they collect and use data, the models that see and act, the world models and simulators they train against and the tools that get a new behavior from the lab onto a fleet of machines. That’s the part of the stack you’ll cover. You’ll track the latest trends in physical AI models and train physical AI systems yourself. You’ll know how a learned policy behaves on a real robot and what a working sim-to-real pipeline looks like. And you’ll be able to tell a system that’s really deployed from a good demo.

What you’ll own & deliver

Edge Research: Working Out What’s Next. Keep up with what’s new in robot learning and work out what it means for the companies we back: VLAs, world models and WAMs, sim-to-real tooling and how to evaluate any of it properly. What you research turns into practical papers, tools, write-ups and advice.

Portfolio Acceleration: Hands-On Impact. Once we’ve invested, you’ll work inside portfolio companies to get robot-learning software from pilot into production: perception, RL policy training and fine-tuning, imitation learning, VLAs/WAMs, world models and simulation, data pipelines, evals and the on-robot inference and fleet tooling that keeps it all working in the field. You’ll work with CTOs and heads of autonomy on results they can measure: task success rate, how often someone has to intervene and how long it takes to stand up a new site.

Find & Win: Technical Due Diligence & Deal Flow. You’ll work with our investment team to look hard at physical AI companies, digging into each company’s autonomy stack, data and roadmap to answer the questions that matter. Is the autonomy learned or scripted? How reliably does it complete the task in messy, real-world conditions, and how often does a person have to step in once you’re past the demo? How long does it take to get a robot working at a new site or on a new task, and does each deployment produce data that makes the next one faster, or is every site a one-off integration? What’s hard to copy: the data, the models, the integration work or the hardware? What you find informs our view of the company, and you’ll talk directly with founders about what our AI Lab could build with them.

We’re looking for

Ideal attributes

• 6+ years building production software; 3+ years leading ML projects, a good share of them on robots or other physical systems
• You’ve put learned perception or policy models on real hardware that ran outside a lab: a warehouse, a factory floor, a field or a road
• Deep knowledge of robot learning: imitation and reinforcement learning, VLA and world models, sim-to-real transfer and the data pipelines behind them
• Solid software engineering: Python, plus enough C++ to read and debug an on-robot stack; ROS 2, real-time and on-device constraints, system design and DevOps
• Enough hardware knowledge to work with mechanical, electrical and controls engineers as a peer, without having to do their jobs
• Comfortable with ambiguity and with taking the lead, a clear writer and speaker

Nice-to-have

• Time at a robotics or autonomy company that went from pilots to a deployed fleet, or on a big-tech or frontier-lab robotics team
• Classical robotics as well as learning: motion planning, state estimation and control
• Published work or real open-source contributions in robot learning (CoRL, RSS, ICRA, LeRobot or similar)
• Experience in logistics, manufacturing, construction, agriculture or energy