Edge AI Engineer (m/w/d)

autonomous-teaming; München; Vor Ort; EUR 60.000 - 80.000; Vollzeit

Zusammenfassung

A defense-tech start-up in Munich seeks an engineer to develop real-time AI solutions for autonomous robotic systems. The ideal candidate will optimize ML models for embedded hardware, addressing challenges like low-latency and debugging. You will work in a high-performing team, focusing on practical implementations in extreme environments. The role emphasizes hands-on experience with CUDA, C++, and TensorRT, and requires a mindset geared towards measurement and validation, aligning with the company’s innovative goals.

Qualifikationen

• Strong experience in CUDA & C++.
• Hands-on work with TensorRT, ONNX, TVM or similar compilers.
• Practical experience with quantization/ pruning/ INT8 / FP16.
• Experience deploying models on Jetson/ embedded GPUs/ ARM / FPGA.

Aufgaben

• Own the full pipeline from model to real-time inference on embedded devices.
• Optimize deep neural networks for Jetson, FPGA, or ARM boards.
• Debug runtime failures and benchmark performance directly on hardware.

Kenntnisse

CUDA

C++

TensorRT

ONNX

PyTorch

Linux

Git

Tools

Embedded GPUs

ARM

FPGA

Jobbeschreibung

Your mission

Build AI that runs in the real world. On real robots. Under real constraints.

At Autonomous Teaming, we build autonomous robotic systems operating in extreme, GPS-denied environments. Our models run fully on edge hardware (Jetson, FPGA, custom boards), with no cloud, no fallback, no excuses.

We’re looking for an engineer who loves hard problems : real-time inference, low-latency pipelines, CUDA kernels, TensorRT graphs, and deploying ML models directly on hardware.

If you enjoy debugging things that only break on the robot, this role is for you.

Missions :

• Own the full pipeline from model to real-time inference on embedded devices:
• Optimize deep neural networks for Jetson, FPGA or ARM boards
• Apply quantization, pruning, distillation to hit strict FPS, power and memory budgets
• Convert & compile models using TensorRT, ONNX, CUDA, C++
• Build ROS nodes integrating optimized perception into the full robotic system
• Debug runtime failures, memory leaks, thermal throttling, kernel-level issues
• Benchmark and validate performance directly on hardware
• Ship models that run reliably in real-world, harsh environments

Your profile

Must-have

• Strong experience in CUDA & C++
• Hands-on work with TensorRT, ONNX, TVM or similar compilers
• Practical experience with quantization/ pruning/ INT8 / FP16
• Experience deploying models on Jetson/ embedded GPUs/ ARM / FPGA
• Comfortable with PyTorch, Python, Linux, Git
• Engineer mindset : measurement→ optimization→ validation

Nice-to-have

• ROS (building nodes, integrating perception stacks)
• Custom accelerators, DSPs or hardware-specific toolchains
• Profilers : Nsight, perf, tegrastats, TensorRT profiler
• Experience in robotics , autonomous systems, aerospace, automotive or defense

Why us?

• You ship code that directly controls real robots
• You work on constrained hardware, where every millisecond and every watt matters
• You solve problems that cloud ML engineers never face
• You own your optimizations end-to-end : from model to field deployment
• You work in a small, high-performing team where ownership is real

If you want a job with clean layers and abstract diagrams, this is not it.

About us

Who we are: Autonomous Teaming is a defence-tech start-up specializing in machine vision solutions. Driven by cutting-edge innovation, our team works on next-generation technologies designed to meet rapidly evolving security challenges.

What we do: We develop systems that enable computers and sensors to operate as coordinated teams, collaborating in real time to counter AI-powered asymmetric threats at scale — including drone swarms and other UXVs. Our mission is to build resilient, intelligent defence capabilities that perform reliably in the most demanding environments.

Who we are: Based in Munich, Berlin, and Toulouse, we are expanding rapidly across Europe with plans to open additional office hubs. We value close, in-person collaboration as the foundation for building complex, high-impact technology, while maintaining flexibility aligned to role and team needs. Our culture is built on ownership, responsibility, and trust — with a shared commitment to growing and building together.