ML Data Engineer - Object Detection & Active Learning

• Design and maintain scalable pipelines to ingest and organize large volumes of time-series camera and sensor data (RGB, IR, acoustic, depth, IMU, and other modalities).
• Own and improve datasets for object detection and classification tasks, ensuring statistical robustness, diversity, and representativeness.
• Build and operate active learning loops to select high-value samples, reduce annotation cost, and continuously improve model performance.
• Write robust preprocessing pipelines using Python, NumPy, Pandas, and augmentation libraries like Albumentations.
• Manage labeling workflows, including QA rules, label consistency checks, vendor coordination, and dataset versioning.
• Collaborate with ML Engineers to fine-tune, train, and evaluate object detection models and feed insights back into the data pipeline.
• Analyze model weaknesses, uncover blind spots, and drive dataset improvements through statistical diagnostics and drift/bias detection.
• Create internal tools to visualize, audit, and analyze dataset quality, diversity, long-tail performance, and failure modes.

• Strong experience with Python and data-processing frameworks (Pandas, NumPy, multiprocessing).
• Familiarity with ETL/ELT pipelines for ingesting, transforming, cleaning, and structuring large video or sensor datasets.
• Experience with data orchestration and lifecycle management for ML workflows (preferably with high-bandwidth video/sensor workloads).
• Understanding of object detection pipelines and formats (COCO, Detectron2, MMDetection, label schemas).
• Experience with active learning, uncertainty sampling, or semi-supervised labeling workflows.
• Familiarity with annotation tools (CVAT, Label Studio) and quality-assurance processes.
• Solid understanding of model evaluation metrics (IoU, mAP, precision-recall, confusion matrix).
• Comfortable working with databases (SQL/NoSQL), file systems, and large-scale image/video dataset organization.
• Ability to collaborate effectively with perception, deployment, and robotics engineers across the product lifecycle.
• Fluent in English, German and/or French are a plus.

Nice to Have:

• Experience with cloud storage systems and MLOps platforms (AWS S3, MinIO, MLFlow, ClearML, Weights & Biases).
• Familiarity with ROS / robotics data formats (bag files, TF trees, sensor_msgs), Docker, or embedded ML stacks.
• Prior work with drone, robotics, or sensor-rich environments, including IR, LiDAR, radar, depth, or audio systems.

Meta:

• Outside-the-box creativity with a blend of conceptual and systematic design thinking.
• High intrinsic motivation, attention to detail, and strong problem-solving mindset.
• Structured, methodical, and reliable execution, even under uncertainty.
• Humble, collaborative, and mission-driven - values collective success over ego.
• High ethical standards and disciplined work ethic.
• Extra-curricular achievements, leadership, or unique projects are a plus.
• NATO-aligned nationality or close ally citizenship is required.
• Successful candidates must obtain security clearance., The world is changing. Exponential technologies are enabling new types of security threats. We are committed to staying ahead by building nimble, scalable, and cost-effective defences. We are looking for passionate developers who are eager to create exceptional products, safeguard our freedom, and strengthen the resilience of democracies.

We are a defence-tech start-up specializing in machine vision solutions. If you have a passion for cutting-edge innovation, and drive to use your skills to create next generation solutions, this is an opportunity for you! What we do: We are developing solutions that enable computers and sensors to collaborate as teams, working together to address emerging security challenges. Our primary mission is to defend against AI-powered asymmetric threats at scale, such as drone swarms and other UXVs. Who we are: Based in Munich, Berlin and Bordeaux/Toulouse we are rapidly expanding across Europe with plans to open more office hubs soon. We embrace a hybrid work culture - valuing the collaborations that happens in the office, while also empowering our team members to work remotely with responsibility and autonomy.