Machine Learning Engineer - LiDAR and Object Recognition (m/f)

As part of a small team, the Machine Learning Engineer (m/f) will develop and deploy models to identify street and infrastructure assets from LiDAR point clouds and associated imagery. The role involves identifying poles, overhead lines, signs, trees, vehicles and other roadside objects, and turning survey data into information for engineering applications.

Responsibilities:

• Work with engineers to define object classes, data requirements, target accuracy and applications
• Develop, train and evaluate models for 3D object detection, classification and point cloud segmentation
• Develop workflows for loading, filtering, aligning, spatially partitioning and georeferencing LiDAR data
• Prepare labelled datasets, coordinate annotation, and track data quality and versions
• Combine LiDAR geometry, camera imagery and geospatial data to improve object recognition and extraction
• Validate models on new locations, sparse data and occluded objects; flag uncertain results for review
• Integrate models into software workflows and export classified point clouds, object locations and GIS data
• Maintain deployed models, document results, and improve accuracy, processing speed and reliability

Potrebno obrazovanje, znanja i vještine

• A degree in computer science, electrical engineering, geomatics, mathematics or a related field, or equivalent practical experience
• 3–5 years of experience in applied machine learning, computer vision or related software development
• Hands-on experience with LiDAR or other 3D point clouds and developing recognition or segmentation models
• Strong Python skills and knowledge of PyTorch or TensorFlow, with experience training, evaluating and deploying models
• Familiarity with Open3D, PDAL, PCL or CloudCompare and data formats such as LAS/LAZ
• Understanding of 3D geometry, coordinate systems, spatial indexing and modern point cloud learning methods
• Experience with data annotation, imbalanced classes and model evaluation, including precision and recall metrics
• Experience with Git, Linux, GPU processing and reproducible data preparation and model training workflows
• Experience with GIS, LiDAR and imagery integration, infrastructure assets, C++ or production ML systems is an asset

Skills and attributes:

• Communication - Communicates clearly with technical and non-technical colleagues
• Independence - Researches solutions and completes tasks with minimal supervision
• Accountability - Takes ownership of tasks and validates results against real-world data

Poslodavac ti nudi

• Flexible working hours
• Career advancement opportunities
• Additional training and professional development
• Remote work opportunities
• Access to the Passport sport platform