Product Manager

About the job

We're looking for a Product Manager to own AI product direction for our intelligent machines — from embodied models and autonomy capabilities through to how they show up in live facilities deployments. You will translate company strategy into clear roadmaps, prioritize across simulation, learning, robotics software, and hardware constraints, and partner with engineering and operations so we ship capabilities that compound in the field.

You will manage products at the intersection of Vision-Language-Action (VLA) stacks, reinforcement learning and policy deployment, simulation and synthetic data, robot hardware, and a production software loop (teleoperation, fleet ops, telemetry, and retraining). You need enough technical fluency to scope tradeoffs, run effective reviews with specialists, and keep delivery aligned with real deployment outcomes — not slide decks.

What you'll do

• Support and own slices of the product roadmap for AI-powered robot capabilities (autonomy features, operator workflows, model-driven behaviors, and supporting platform surfaces)
• Help define and prioritize work across VLA training and deployment, RL policy rollout, simulation assets, data/telemetry pipelines, and field evaluation — balancing research ambition with shippable milestones
• Partner with simulation, RL, perception/autonomy, software platform, and electrical/embedded teams to track dependencies (environments → datasets → models → safe deployment → fleet learning)
• Write clear PRDs, acceptance criteria, and release notes that reflect embodiment differences, latency/safety limits, sim-to-real risk, and serviceability in deployed facilities
• Run discovery with facilities operators and internal ops: turn tasks (cleaning, inspection, monitoring, maintenance support) into structured requirements and measurable field outcomes
• Support product rituals: roadmap reviews, release checklists, and failure-mode reviews that feed prioritization
• Track adoption, reliability, and learning flywheel metrics (deployment coverage, intervention rates, eval regressions, data volume/quality for retraining)
• Communicate status and tradeoffs to leadership and engineering leads with crisp written updates

What we're looking for

• 1–2 years of product management experience (or equivalent: associate PM, product analyst, or technical program coordinator role with clear product ownership)
• Track record shipping or co-shipping technical products — internships, startup roles, or internal tools used by engineers or operators count
• Ability to work across multiple engineering disciplines without being a hands-on ML researcher
• Working knowledge of modern robot learning stacks: VLA or multimodal policies, RL/imitation/offline RL basics, evaluation harnesses, and sim-to-real considerations
• Familiarity with robotics simulation workflows (e.g. Isaac Sim/Lab, MuJoCo, synthetic data, domain randomization) and how they gate model training
• Understanding of deployed robot systems: teleoperation, fleet monitoring, versioning of models/policies, rollback, and operational safety constraints
• Comfort with hardware and embedded realities as product boundaries (power, sensors, compute, reliability, field service) even if you are not an EE
• Strong written communication, good judgment on scope, and credibility in technical design and standup reviews
• High agency, bias to clarity, and ability to thrive in a pre-seed / early-stage environment with evolving scope

Nice to have

• Internship or full-time experience in robotics, embodied AI, ML platforms, dev tools, or hardware-adjacent software
• Exposure to facilities management, hospitality, airports, or other high-variation commercial environments
• Coursework or project experience with data flywheels: telemetry → curation → retraining → staged rollout
• Familiarity with ROS / ROS 2 concepts, edge inference, or MLOps for on-robot deployment
• Side projects or thesis work touching manipulation, mobile robots, or simulation for learning
• User research with operators, B2B pilots, or enterprise deployment playbooks

Who you are

• You want AI products that survive contact with messy facilities, not demos that only work in the lab
• You think in systems: models, sim, hardware, software platform, and operators as one product surface
• You are ambitious, collaborative, and comfortable asking questions until tradeoffs are clear
• You care about compounding advantage — every deployment should make the next one easier to ship and safer to run
• You want to grow into defining what intelligent machines mean as a product category in Southeast Asia and beyond

What we are looking to build

AI product lines that connect our full stack — embodied models (VLA and RL-adjacent policies), simulation and evaluation infrastructure, deployment platform, and robot hardware — into coherent capabilities for facilities operations, with roadmaps and release discipline that turn field data into durable product and model advantage.