Career Break & Personal Development
2025 - Present
Sabbatical & focused learning
- Intentional break after several years in automated driving — traveling and going deep on AI engineering, LLMs, and agent-based workflows.
Thomas Blank's Resume
Thomas Blank
Senior Software Engineer · ML Infrastructure
About
I'm a senior software engineer specializing in ML infrastructure. At Bosch, I co-created xflow — a modular training and deployment framework adopted by 10+ teams. I care about clean architecture, reproducible systems, and making ML work reliably outside the lab.
Work Experience
Bosch, Level 4 driving project
2021 - 2025
Senior Software Engineer - Machine Learning
- Architected and co-created xflow, a modular ML training and deployment framework now used by 10+ teams across Bosch.
- Designed and optimized data and model pipelines, reducing training iteration cycles from days to hours and cutting associated compute costs.
- Contributed to migrating the perception stack from single-view CNNs to a multi-view spatio-temporal architecture, enabling detection of edge cases the previous system couldn't handle.
- Engineered reproducible training and inference workflows with Docker, AzureML, and CI/CD integration.
- Collaborated closely with corporate research to bring state-of-the-art models into production, including ONNX/TensorRT optimizations for embedded deployment.
- Established team-wide best practices for code architecture and ML pipelines; owned architecture/code reviews and ran technical interviews for new engineers.
Bosch & Daimler Collaboration Project Athena
2018 - 2021
Software Engineer - Machine Learning
- Developed and maintained model pipelines for object detection and tracking on automotive datasets.
- Deployed and benchmarked models with TensorRT, achieving real-time inference on embedded GPUs.
- Improved training reproducibility and evaluation by automating dataset versioning and experiment tracking.
YAZIO
2014 - 2017
Full Stack Web Developer
- Built and maintained high-traffic web APIs (Symfony2, Redis, PostgreSQL) serving >1M monthly users.
- Automated deployments with Ansible and implemented backend monitoring for reliability and speed.
Education
Master's Degree in Applied Computer Science, FH Erfurt
2014 - 2018
Thesis: Machine Learning based analysis of complex sign plates (object detection & semantic segmentation)
Bachelor's Degree in Applied Computer Science, FH Erfurt
2011 - 2014
Skills
- Python
- C++
- Bash
- PyTorch
- TensorFlow
- scikit-learn
- ONNX
- TensorRT
- Linux
- Git
- Docker
- Azure ML
- CI/CD
- Distributed Training
- Data Pipelines
- Experiment Tracking
- GitHub Actions / Azure DevOps
Languages
- German—Native
- English—Fluent (C1)
Side projects
NextFlight
Flight-debrief app for paraglider and hang-glider pilots — analyzes uploaded IGC/GPX tracks entirely in the browser to produce concrete takeaways for the next flight.
- Next.js
- TypeScript
- Browser-Only
Relay Mail
Fork of Thunderbird for Android with biometric app-lock (fingerprint / device credential). Built to scratch my own itch, in daily use.
- Kotlin
- Android
- Biometrics