Applied AI Engineer, Robotics

Mistral AI is seeking an Applied AI Engineer focused on robotics deployment. You will take state-of-the-art AI models and make them work on real robots, in real environments, for real customers. This is a hands-on role at the intersection of machine learning and physical systems: you will own the path from model to deployed autonomy, working alongside robotics engineers, ML researchers, and systems engineers to deliver end-to-end autonomous solutions across diverse use cases.

What you will do

• Integrate AI models with robotic hardware, sensors, and embedded systems, bringing modern perception and language-driven capabilities onto physical platforms.

• Improve the robustness, reliability, and safety of robotic systems operating in real-world, unstructured environments.

• Test and validate algorithms in simulation and in real-world deployments, moving fluidly between the two.

• Analyze field data to improve model performance and system reliability, closing the loop between what happens on the robot and what happens in training.

• Bring up, tune, and debug the autonomy stack — localization, mapping, navigation, and perception — on real hardware, from sensor calibration to real-time performance.

• Develop and tune robot navigation behaviors — path planning, obstacle avoidance — so systems hold up in dynamic, cluttered, and partially observable environments.

• Diagnose failures end-to-end: mapping drift, localization dropouts, perception edge cases, timing issues — and fix them.

• Collaborate with robotics engineers, ML researchers, and systems engineers to deliver complete autonomous solutions for customer use cases, from prototype through production deployment.

• Build and maintain the tooling needed for deployment, monitoring, and continuous improvement of deployed systems.

• Bachelor's, Master's, or PhD in Mechatronics, Robotics, or a relevant discipline (e.g. Computer Science, Electrical or Mechanical Engineering).

• Fluent in English with excellent communication skills.

• Strong software engineering in Python and/or C++: clean, readable, high-performance code; comfortable working in and improving production codebases.

• Knowledge of robotics frameworks such as ROS/ROS2.

• Experience with robot perception and state estimation: SLAM, localization and mapping, path planning, sensor fusion, or computer vision.

• Solid mathematical fundamentals — geometry, probability, and estimation — and the judgment to know when a hand-crafted algorithm beats a learned one.

• Familiarity with simulation tools and robotics development environments (e.g. Isaac Sim, Gazebo, or similar).

• Strong problem-solving skills and ability to work in interdisciplinary teams.

• Comfortable with the messiness of the real world: hardware quirks, edge cases, and field surprises don't scare you.

• Low-ego, collaborative, and eager to learn.

• Doesn't need roadmaps. Ships.

Nice to have

• Experience deploying ML models on embedded or edge hardware (GPU/NPU inference, quantization, real-time constraints).

• Hands-on experience with sensor modalities such as LiDAR, depth cameras, or IMUs, including calibration and time-synchronization.

• Exposure to VLM/VLA models or other foundation-model approaches for robot perception and control.

• Experience with navigation in GPS-denied or otherwise challenging environments, or with visual-inertial odometry.

• Working knowledge of established autonomy tooling — e.g. Nav2, SLAM Toolbox, Cartographer, RTAB-Map, or AMCL — and knowing when to build versus reuse.

• Experience with behavior trees, dynamic costmaps, or fleet-level navigation.

• Contributions to open-source robotics or ML projects.

• Experience with safety-critical or regulated deployment environments.

• Track record of success through personal projects, professional projects, or academia.