MLOps Engineer (m/f/d) – Embodied AI & Robotics Platform

Das ist der Job

My client is a highly innovative, venture-backed robotics pioneer based in Munich's thriving deep-tech hub.

Darum lohnt es sich

They will collaborate heavily with AI Researchers, Core Robotics Software Engineers, and Cloud Infrastructure teams to ensure that the company’s moving fleet of robots can reliably see, reason, and act.

Modern Work Culture: Flexible hybrid working arrangements, 30 days of annual paid vacation, and access to state-of-the-art laboratory and testing spaces in Munich.

Relocation Support: Comprehensive assistance with visa processing, bureaucratic onboarding, and relocation expenses for international talent moving to Munich. #J-18808-Ljbffr They bridge the gap between advanced artificial intelligence and the physical world by building state-of-the-art robotic platforms designed to automate complex tasks in dynamic, real-world environments.

By integrating next-generation hardware with advanced Computer Vision, Spatial Intelligence, and Embodied AI, they empower industries to run more fluidly, efficiently, and safely.

Role Overview The company is seeking an engineering-focused MLOps Engineer with a foundational background in Applied AI to take ownership of the machine learning lifecycle pipeline.

This is not a typical cloud-SaaS or FinTech MLOps role; the engineer will solve the unique challenges of Edge MLOps – taking complex perception and vision models from cloud-based experimentation and deploying them directly onto resource-constrained physical robotic hardware.

The ideal candidate sits at the intersection of Machine Learning, Infrastructure Engineering, and Hardware-Software Integration. Key Responsibilities Pipeline Automation: Design, build, and maintain robust, end-to-end ML orchestration pipelines (from data ingestion and synthetic data generation to distributed multi-GPU training and model testing).

Hardware Optimization & Edge Deployment: Develop automated compilation, quantization, and pruning workflows to optimize deep learning models (e.g., Computer Vision, Transformer-based architectures) targeting edge-compute accelerators like NVIDIA Jetson or custom automotive/robotic SoCs.

Data Fleet Infrastructure: Engineer high-throughput, multi-modal data loops capable of handling massive streams of sensory data (camera frames, point clouds, telemetry) and establish smart data‑curation methods to upload edge‑edge failures for retraining.

Lifecycle & Experiment Management: Set up and enforce rigorous experiment tracking, model registry, versioning, and approval gates using tools like MLflow or Weights & Biases to guarantee absolute reproducibility.

Continuous Monitoring & Drift Detection: Build observability stacks and dashboards (e.g., Prometheus, Grafana) to track real-time model inference degradation, sensor drift, and physical environment anomalies out in the field.

CI/CD Integration: Construct reliable infrastructure-as-code (IaC) and containerization frameworks to push seamless Over-the-Air (OTA) model and firmware updates securely to a fleet of physical robots.

Profile & Requirements Education: University degree (B.Sc., M.Sc., or Ph.D.) in Computer Science, Robotics, Electrical Engineering, Data Infrastructure, or a comparable technical field. Experience: 2 to 5 years of professional experience in MLOps, DevOps, or Machine Learning Infrastructure engineering roles.

Software Skills: Production-grade Python skills are mandatory. Comfort working in Linux environments and a familiarity with C++ or hermetic build systems (e.g., Bazel) is highly beneficial.

AI/ML Foundations: Solid baseline understanding of Machine Learning and Deep Learning concepts, specifically with hands‑on exposure to frameworks like PyTorch , OpenCV, or Hugging Face Transformers.

Cloud & Infrastructure: Proven experience managing production workloads in containerized environments via Docker and Kubernetes (including experience configuring accelerator node pools/GPUs). Strong familiarity with AWS, GCP, or Azure infrastructure and Terraform.

Nice-to-Haves: Exposure to the Robot Operating System ( ROS/ROS2 ), robotics data serialization/logging tools (like MCAP or Foxglove), or model optimization runtimes (TensorRT, ONNX Runtime). What the Company Offers Deep Tech Impact: The chance to work directly on physical robotic platforms and watch code directly translate into hardware mechanics.

Highly Competitive Compensation: Attractive base salary paired with equity package structures and performance-driven bonuses.