Senior Data Scientist – Python, ML & Signal Processing
DATATRONiQ is a deep-tech startup from Germany for Industrial IoT and Edge AI. As Senior Data Scientist you train machine-learning models on real machine and sensor data — at customers worldwide, from mid-market manufacturers to DAX corporations. Anomaly detection, predictive maintenance, and real-time quality monitoring. The data is real, the machines are on running production lines, and your models help decide whether a line stays on plan or slips into unplanned downtime. If you'd rather see your models in production than optimize them in notebooks — let's talk.
- Location
- Stuttgart, Ulm or Berlin
- Level
- Senior · 3+ years experience
- Stack
- Python · PyTorch · scikit-learn · ONNX · GPU
About the role
You own the full data-science lifecycle — from exploration through feature engineering and model training to deployment and validation, depending on the customer situation on an edge gateway, an on-prem server, or in the cloud. The time series arrive from industrial controllers over OPC-UA and MQTT; feature engineering here means signal processing on noisy sensor data, not reshuffling clean table columns. Your models get quantized, exported to ONNX, and run where production needs them — with everything that implies for model choice, latency, and memory footprint.
Your stack: Python, PyTorch or scikit-learn, ONNX for edge deployment, and common MLOps tooling for versioning and reproducibility.
We work in a small, tightly coordinated team. Code reviews and pair programming are a fixed part of the week. Every Friday we share interesting finds from the web and new tools in our show-and-tell sessions. As a data scientist you work closely with data engineers and backend developers: you build the models, they build the pipelines, and together you get both into production.
What are your responsibilities?
- You train ML models for anomaly detection, predictive maintenance, and quality monitoring, and validate them on real production data and time series from industrial controllers.
- You do feature engineering on noisy sensor and machine signals (OPC-UA, MQTT, MES exports), including signal processing and filtering.
- You deploy models where the customer needs them — edge gateway, on-prem server, or cloud: quantization, ONNX export, sometimes with constrained CPU and RAM resources, tuning and in-field monitoring.
- You work closely with data engineers and backend developers so models ship reliably into our production pipelines — not as a notebook prototype.
- You measure models not only by their F1 or AUC scores but by what they deliver in production: fewer unplanned outages, higher output, fewer errors.
- You contribute actively to the product roadmap and technical decisions — we expect opinions, not just ticket churn.
What is your qualification?
- Completed degree in Data Science, Computer Science, Mathematics, Physics, or a related field.
- At least three years of very hands-on experience with Python, common ML frameworks (PyTorch, scikit-learn, etc.), and model deployment into production environments.
- Experience with time-series analysis and signal processing — we don't expect PhD-level depth, but you know why a naive MLP fails on noisy industrial signals.
- Working knowledge of MLOps: versioning for models and data, reproducible pipelines, tests for ML code.
- Strong written and spoken English.
- You can make a technical case and defend it in the team — including against a majority, when you have strong arguments.
- Bonus: experience with edge deployment (ONNX, TensorRT, quantization), industrial protocols (OPC-UA, MQTT), or LLMs for chat and agentic tasks.
What can you expect?
- Real end-to-end ownership: from scoping the data capture on the production floor through the pipeline to model inference — on edge gateway, on-prem server, or in the cloud.
- The team decides architecture, tooling, and model strategy.
- Agentic tooling in daily work: Codex, Claude Code, and new development practices — we try them early and use what works.
- Mostly on-site in Stuttgart, Ulm, or Berlin — Industrial IoT projects for customers worldwide, from mid-market manufacturers to DAX corporations, on real production data.
Did we spark your interest?
This is a unique and exciting opportunity to work on a product that has the potential to make a significant impact in the industrial manufacturing sector. DATATRONiQ platform is built with TypeScript, Golang, Python, Scala, C++ and deployed as microservices on various Cloud environments and on-prem virtual machines using Kubernetes.
If you are passionate about using your skills to solve challenging problems and make a difference, we really would like to talk to you.
Get in touch with us at work@datatroniq.com. We are looking forward to your application!