Internship: Machine Learning (m/f/x) (Geospatial AI)

Location: Munich, Germany; hybrid workplace

Salary: €2,410 per month

Company

hirify.global develops SaaS solutions that improve transparency, efficiency, and trust across global value chains, using AI to support human rights, ecological responsibility, and sustainable corporate governance.

What you will do

• Train, fine-tune, and evaluate deep learning models on satellite imagery using PyTorch.
• Analyze existing models and maps, identify failure cases, and improve their performance.
• Implement recent computer vision and remote sensing research in working code.
• Convert prototypes into production-ready code for large-scale geospatial pipelines.
• Own projects from data preparation and training through evaluation and deployment.
• Write tested, efficient, maintainable, and well-documented Python code and communicate technical ideas clearly.

Requirements

• Late-stage Bachelor's or Master's studies in Computer Science, Electrical Engineering, or a related field.
• Strong understanding of machine learning concepts and algorithms.
• Hands-on experience building and training deep learning models with PyTorch.
• Solid Python engineering skills, including Git and Unix-based systems.
• Ability to read research papers and implement their methods.
• Excellent English communication skills and availability for hybrid work in Munich.

Nice to have

• Computer vision research experience.
• Experience with satellite imagery or geospatial data, including multispectral imagery, GeoTIFFs, GeoJSON, rasterio, or GDAL.
• Experience scaling training or inference with GPUs, distributed training, or cloud environments.

Culture & Benefits

• Six-month fixed-term, full-time internship.
• Work on real-world geospatial AI and remote sensing problems related to deforestation-free supply chains.
• Exposure to the full machine learning lifecycle, from data and training to evaluation and deployment.
• Flexible hybrid work options from a modern Munich office.
• Open team culture and flat hierarchies.
• Opportunity for successful models to reach more than 2,500 customers in production.