3D ML Engineer

Description

You will be working on an advanced R&D project focused on 3D data processing, geometric analysis, and machine learning for spatial datasets obtained from photogrammetry (mainly DJI drones). The goal of the project is to develop a module capable of :

importing and interpreting 3D data from industry-standard tools (point clouds or mesh),

detecting and segmenting buildings from complete 3D models,

simplifying geometry into a lightweight vector representation for further technical processing.

Your work will span various phases of end-to-end delivery : understanding the business context, analyzing technical requirements, proposing algorithms, implementing 3D processing and ML components, and supporting integration with production systems. You will collaborate closely with developers responsible for 3D → 2D pipelines and technical drawing generation.

This is an opportunity to contribute to a high-impact module that will be showcased to clients and industry partners.

Requirements

Primary qualifications :

Master of Science in computer science, mathematics, physics, engineering, or a related field.

Strong experience with 3D data processing – point clouds, mesh models, spatial data formats (.las, .ply, .obj, .xyz, .e57).

Practical knowledge of libraries such as Open3D, PCL, PDAL, CGAL, MeshLab or similar.

Hands-on experience in segmenting point clouds

Experience with machine learning / deep learning for 3D data, including models such as PointNet / PointNet++, KPConv, MinkowskiEngine, SparseConvNet, or CNN-based approaches on 2D projections.

Ability to build and train ML models for geometry identification, segmentation, and classification.

Experience with geometric algorithms : plane detection (RANSAC, Hough transform), edge detection, angle estimation.

Understanding of mesh simplification techniques (e.g., Quadric Error Metrics).

Experience working with photogrammetric outputs from tools like DJI Terra, Pix4D, RealityCapture, or Agisoft Metashape.

Strong proficiency in Python (required); C++ is a plus.

Familiarity with scientific and ML libraries : NumPy, SciPy, PyTorch / TensorFlow, scikit-learn, Open3D.

Experience processing large-scale 3D datasets, optimizing pipelines for GPU / CPU performance.

Ability to maintain high-quality, test-covered code, including unit, integration, and load tests.

Experience using Linux, command line tools, Git, Docker.

Strong analytical thinking and problem-solving skills.

Ability to work in a team and document work clearly.

Willingness to learn and develop skills continuously.

Communicative English — minimum B2 level.

It is great if you have :

Experience with DevOps / MLOps tools (Docker, Kubernetes, MLFlow, KubeFlow, DVC).

Knowledge of deep learning frameworks : PyTorch or TensorFlow.

Experience with additional data science libraries (OpenCV, scikit-image, plotly, seaborn, xgboost, lightgbm).

Experience in computer vision or photogrammetry-heavy projects.

Understanding of CAD / BIM standards and building geometry fundamentals

Experience with 3D data visualization tools (Open3D viewer, MeshLab, Potree), or WebGL / Three.js.

Skills in algorithmic optimization and performance tuning.

Experience with cloud platforms (AWS, GCP, Azure).

Salary :

20 000 - 26 000 PLN + VAT (B2B)

We offer you :

Working with modern machine learning and 3D processing technologies.

Annual self-development budget.

Participation in a variety of innovative, challenging projects.

Internal workshops and knowledge-sharing sessions.

Personal branding opportunities (articles, conference speaking, leading workshops).

Flexible working hours.

Remote work possibility.

Friendly atmosphere and supportive team environment.

Chillout room / free beverages / team & company events (for on-site visits).

MultiSport card.

LuxMed healthcare package.