Breaker develops on-robot AI agent software that lets military operators control fleets of autonomous air, land, and sea systems using natural-language voice commands.
PERCEPTION ENGINEER — LOCALISATION & SENSOR FUSION - SYDNEY TL;DR - Own and evolve Breaker's 3D localisation pipeline — turning 2D detections into accurate real-world coordinates using passive sensors only - Deep-stack work in camera-based localisation, sensor fusion, and real-time data processing on edge hardware - Looking for a mid-to-senior engineer who has deployed production localisation systems on moving robotic platforms - You'll build the metrics, tooling, and processes to continuously improve localisation accuracy across diverse hardware configurations - Must have working rights in Australia and be willing to participate in regular field testing - Join an exciting startup backed by globally recognised investors at the bleeding edge of physical AI ABOUT BREAKER The way humans use robots is broken. Modern problems demand more robots than we have operators. Every drone, ground vehicle, and maritime system requires dedicated training, manual control, and constant oversight. One operator per robot. One pilot per mission. Breaker's AI agent breaks this constraint. Our technology turns any robot into a truly autonomous, self-organizing teammate. Operators command and query teams of robots through natural language conversations over the push-to-talk radios they already carry, no laptops, no controllers, no additional gear. Instead of manually flying search patterns across three different screens, you say, "survey this area and flag anything unusual." The robot team figures out how to divide the task, coordinate their movements, and report back what matters. We're fundamentally changing the operator-to-robot ratio. Small teams become force multipliers. Our software deploys directly onboard each robot, enabling real-time, intent-driven control even in contested environments with limited bandwidth. We're solving problems most AI companies never touch: sub-second inference on edge hardware with strict latency, power, and connectivity constraints. We're backed by some of the best investors globally and growing our team across Austin, Texas, and Sydney, Australia. We're a small team of experienced engineers moving fast on technology that will define how humans and machines work together for decades to come. Join us if you want to help build the robots we were promised 🤖 ABOUT THE ROLE One of the most critical outputs of Breaker's product is putting accurate points on a map — taking 2D detections from camera systems and resolving them into precise 3D world coordinates. This is a hard, fascinating problem: we do it without LIDAR or active ranging, using passive sensors on moving platforms in unpredictable environments. You'll own this pipeline end-to-end. This is a greenfield ownership opportunity. You'll take an existing localisation system and make it yours by building the measurement infrastructure to quantify performance, researching and implementing new techniques, and expanding existing capability into areas like tracking moving objects from moving platforms and generalising across diverse camera hardware. You'll define the contracts between your system and the teams that feed into it (computer vision, hardware, field operations) and build the data workflows that turn real-world testing into continuous improvement. You'll work at the deepest level of our software stack, where real-time sensor data meets geometric reasoning. If you're the kind of engineer who gets energy from digging into a dataset to find where localisation breaks, running experiments to prove a new approach works, and building the automation to measure it all. KEY RESPONSIBILITIES - Own the accuracy and performance of Breaker's 3D localisation pipeline, from 2D bounding boxes in, lat/long/alt out - Build metrics, ground truth collection processes, and automated evaluation tooling to measure and track localisation performance - Research, prototype, and deploy new localisation and sensor fusion techniques to improve accuracy and robustness - Extend the system to track moving objects from moving platforms - Generalise the localisation pipeline to deploy reliably across different camera systems and hardware configurations - Define interface contracts with adjacent systems like what the CV pipeline must deliver, what camera hardware must provide - Process and analyse field data to identify failure modes, validate improvements, and feed findings back into the development loop - Design and specify field test scenarios (CONOPs-style) to stress-test specific localisation behaviours - Participate in regular field testing, including training field teams on data collection requirements - Conduct software-in-the-loop and hardware-in-the-loop testing and experimentation ABOUT YOU REQUIRED SKILLS & EXPERIENCE - Demonstrated experience with camera-based localisation on moving robotic platforms. Monocular or stereo, not LIDAR-dependent - Strong sensor fusion fundamentals. You understand timing, calibration, and the realities of combining data from multiple real-time sensor streams - Experience processing real-time data on constrained or edge-deployed hardware - Proven ability to dig into robotic datasets, identify where systems fail, and systematically troubleshoot localisation problems - Familiarity with optimisation backends (Ceres, GTSAM, or similar) for geometric inference - Proficiency in Python and/or C++ - Comfortable working in Linux environments - Have deployed a production system, not just research prototypes or thesis projects - Must have working rights in Australia - Willing to participate in regular field testing HIGHLY VALUED EXPERIENCE - GPS-denied navigation or localisation work - Photogrammetry or large-scale imagery processing pipelines - Robotic manipulation with visual servoing or camera-based feedback - Experience with gimbaling camera systems and their associated complexities - Understanding of machine learning s