We are looking to add a Senior Medical AI & Clinical Data Engineer (m/f/d) to our growing team.
We develop SeeCure, a mobile ophthalmic diagnostics platform (SaMD), in clinical partnership with the LMU eye hospital in Munich. Core development is funded by the German Federal Ministry of Research (BMFTR). As a next step, we are creating a Ophthalmic foundation model. We are developing a new AI-based product for forecasting glaucoma progression and detecting rare optic neuropathies. In this role, you will be our first AI hire and technical lead for this product, instrumental in its architecture, tooling, and building the team from day one.
Your responsibilities
- Own the end-to-end data pipeline: ingestion (DICOM, vendor exports, FHIR), quality rules, patient-level linkage and curation of perimetry, OCT and longitudinal clinical records inside our on-premises trusted research environment.
- Define the data specification, codebook and annotation guideline together with the eye hospital's bioinformatician and clinicians; run the feedback loop on every data delivery.
- Design and train the models for progression forecasting and atypia detection, starting from an initial approach (self-supervised sequence encoding for longitudinal visual-field data, continued pretraining of an open vision transformer such as DINOv2 on OCT volumes, and multimodal fusion with a time-aware head), which you are expected to validate and adjust based on the available data.
- Build reproducible training and evaluation infrastructure (PyTorch, Lightning, MLflow, DVC, ONNX) and a pre-registered evaluation protocol with patient-level splits, calibration and subgroup analysis.
- Provide technical guidance to the ML Engineer and working student joining in 2027, and collaborate closely with our external medical AI advisor.
- Take ownership of the technical content underlying our conformity documentation under IEC 62304, ISO 14971, and the EU AI Act. All architectural decisions, dataset selection criteria, and evaluation results must be traceable and defensible to a notified body. You will get support with drafting; the technical accuracy and traceability of the content are your responsibility.
- Publish, contribute to public benchmarks, and report progress to funding bodies.
Your profile
- M.Sc. or Ph.D. in computer science, data science, medical informatics, physics, or a related field, with 5+ years of hands-on machine learning experience and a proven track record of taking models from prototype into production, ideally on medical or other sensitive data.
- Strong command of PyTorch and Python, with solid software engineering practices (Git, testing, CI, containers), and experience owning an ML system end to end, from the data pipeline through deployment.
- Practical experience with self-supervised learning or vision transformers (e.g. DINO-family, masked autoencoding), and with DICOM, FHIR, or clinical time series data.
- Comfortable owning real technical direction: this is a zero-to-one build, and you are expected to set the architecture, the standards, and the bar for the engineers who join after you.
- Experience working under a regulated or quality-managed development process (ISO 13485, IEC 62304, or equivalent), since you will own the technical substance of our conformity documentation.
- Credible across audiences: able to defend a modeling decision to a notified body, explain a limitation to a clinician, and justify a data-handling choice to a data protection officer.
- Fluent in German and English. Domain knowledge in ophthalmology or neurology, MLOps experience, and a publication record are a plus.
What we offer
- The lead technical role in developing an Ophthalmic foundation model. An AI product, working directly with clinicians and an established medical AI advisor.
- Dedicated on-premises compute: a GB10 (Blackwell) development node and a dedicated Blackwell training workstation.
- Flexible working hours, with hybrid arrangements for development work and on-site presence required for work involving patient data within the trusted research environment.
- A permanent contract with fair compensation.
- Participation in publications and patents.
- Our founding team has already brought regulated software and a VR-based clinical endpoint to market.
Apply with your CV and portfolio links (GitHub, publications) to [email protected].
Full-time · Start December 2026 or by arrangement
https://goodly-technologies.de/careers.html