Senior Deep Learning Engineer (m/f/d)

Your Mission: What You'll Do

* Advance our deep learning models through research iteration: develop and test architectural improvements, training objectives, and optimization techniques. You will own the loop from idea → experiment → conclusion → next iteration.

* Build rigorous evaluation and benchmarks: define evaluation sets, establish clear metrics (precision, recall, accuracy, calibration), and create repeatable benchmark runs so improvements are measurable and comparable over time.

* Own monitoring of model quality: set up monitoring for model performance and data shifts, define alerting signals, and build lightweight reporting that makes regressions visible early.

* Partner cross-functionally to turn findings into impact: work with data/engineering teams to improve datasets and labeling strategies, and with product/ops stakeholders to align on what “good” looks like in practice.

Your Profile: Qualifications & Requirements

* MSc in Computer Science or a related field with 4+ years of applied deep learning experience, or a PhD — paired with a proven track record of taking research from idea to working system.

* Strong understanding of modern neural architectures, with demonstrated depth in transformer architectures and practical experience improving and adapting them (e.g., attention variants, efficiency improvements, robustness, training stability).

* Strong Python deep learning stack experience (e.g. PyTorch, Tensorflow), including training pipelines, experimentation discipline, and reproducibility.

* Solid experience with experiment tracking and model evaluation tooling (e.g. Weights & Biases or similar), and a strong bias for measurement-driven progress.

* Fluency in English, German is a plus

Nice to have

* Experience with domain adaptation and evaluation in imbalanced settings (rare events, high cost of misses/false alarms).

* Familiarity with deployment-adjacent concerns: model packaging, performance constraints, and continuous evaluation in changing real-world conditions.

* Experience working with sensor/time-series or industrial data, where edge cases and dataset shifts are the norm.

* Experience with agentic AI development workflows to speed up experimentation (analysis, ablations, test scaffolding) while maintaining careful review and scientific rigor.