Machine Learning Engineer — Industrial Robot Vision

Chiara Lombardi

Contract · Mid-level · Berlin

This role is no longer accepting applications

Similar open roles you can apply to now:

About the role

We are a robotics division of a Japanese precision engineering conglomerate. Our robots are deployed in electronics assembly, pharmaceutical packaging, and semiconductor handling applications globally. The common requirement across all of these is precise manipulation in controlled environments — a robot that places a chip on a PCB needs sub-millimetre accuracy and zero tolerance for misidentification. Our current machine vision systems are rule-based, developed over many years by our in-house vision engineers, and they work. They work until the product changes, at which point the manual tuning process to adapt them is slow and expensive. We have been evaluating ML-based vision approaches for several years and have concluded that the technology is mature enough to deploy selectively — not as a replacement for our rule-based systems in high-criticality tasks, but as a complement for the flexible inspection and object recognition tasks where the rule-based approach struggles with product variation. We are hiring a contract ML engineer for a 10-month engagement to work within our European R&D centre in Stuttgart on integrating ML-based vision into two specific automation programmes. The work requires German or English — the team works in both.

Responsibilities

Develop and validate ML-based vision models for object identification and inspection for two active automation programmes
Conduct systematic evaluation against existing rule-based system benchmarks with statistical rigour
Optimise inference performance for real-time operation within the cycle time constraints of the assembly process
Produce detailed technical documentation for every experiment, model decision, and evaluation result — this will be reviewed by senior R&D engineers
Deliver a structured handover with a clear recommendation on production readiness at the end of the engagement

Requirements

5+ years of ML engineering with a focus on computer vision for industrial automation, quality inspection, or precision robotics
PyTorch for training object detection and classification models on constrained industrial datasets — experience with small-data regimes and data augmentation strategies is directly relevant
Python at a professional engineering standard with the discipline to write code that survives handoff to a team that has been here for twenty years
MLflow for experiment tracking and model versioning — we need complete reproducibility of every result
Scikit-learn for baseline model development and evaluation — classical methods often win on small industrial datasets and you should know when to use them
Docker for reproducible environments deployable to our on-premises R&D compute infrastructure
Experience with precision evaluation: understanding the cost asymmetry between false positives and false negatives in manufacturing quality contexts

Benefits

€680 – €840 per day (contract rate)
10-month initial engagement
On-site in Stuttgart with flexibility for remote work during non-lab phases — we will cover relocation costs for the on-site period
Access to proprietary precision component datasets not available in academic settings
Work in an R&D environment where engineering standards are exceptionally high