At deeplify, we’re building the first AI-native asset integrity co-pilot for critical industrial infrastructure. We turn inspection data from pipelines, chemical plants, ships, and bridges into real-time, risk-based maintenance decisions. We combine a digital inspection platform with proprietary deep-learning models and an evolving agentic AI system that learns from asset integrity engineers. This shifts asset integrity from slow, analogue, document-driven processes to a proactive, software-defined, and increasingly autonomous system.

Tasks

We are looking for an exceptional ML engineer working student to help us solve some of the hardest applied machine learning problems in industrial inspection — from weld defect detection and corrosion analysis on radiographic data to future UT-based systems and long-term corrosion prediction.

This is not a narrow research role. It is about solving hard end-to-end real-world problems: turning messy industrial data into reliable production systems.

• Deep learning models for weld defect detection and corrosion analysis on radiographic and ultrasonic data
• Managing external labeling teams
• Training, evaluation, and experiment tracking workflows
• Production inference pipelines
• Support an exciting research project

Requirements

• Strong hands-on ML engineering skills
• High ownership: you take responsibility, drive things forward, and do not wait to be told every next step
• High urgency: you move fast, care about execution, and know how to create momentum
• Excited by messy, difficult, real-world problems with no obvious solution
• Comfortable working across data, models, infrastructure, and deployment
• Bonus: experience in computer vision, MLOps, production ML, imaging, or sensor data

Benefits

• Work on technically ambitious problems with real industrial impact
• Build end-to-end ML systems, not just models in isolation
• Help lay the foundation for a scalable internal ML platform
• Be part of a team tackling long-term challenges like corrosion prediction, a genuinely hard problem with significant upside
• Well above average working student compensation