Software Engineer - Python

We build the tooling that gets pathology data from a scanner into a GMP-regulated AI pipeline — de-identification, quality checks, dataset management, metadata handling for whole slide images at tens of thousands of pixels per side. The work is Python- heavy, scientifically grounded, and deliberately unglamorous: CLI tools, data validation, lightweight UIs, and pipelines that have to be correct. If you’ve built tools to solve real problems in a quantitative field, you’ll recognize this kind of work. The role You’ll work embedded with a small computational pathology team, picking up tasks day-to-day as a remote contractor. The work spans the full tool lifecycle: gathering requirements from scientists and pathologists, designing and building the solution, writing tests, and maintaining what you ship. Concretely, you’ll be working on things like: - Data ingestion and metadata pipelines for whole slide images - De-identification and integrity checking of WSI files - Dataset splitting and stratification tools - Lightweight review and management UIs - CI, packaging, and test coverage for the above This is not a backend or frontend role. There’s no microservice architecture to maintain. The problems are scientific and data-oriented; the solutions are focused,deployable Python tools. We integrate AI assistance heavily into our development workflow — from requirements exploration to code review. Comfort working alongside AI tooling is a plus; experience shaping how a team uses it is even better. What we’re looking for Essential - Strong Python — you write clean, testable, well-documented code without being prompted - Comfort working with data: tabular data, file-based pipelines, validation, transformation - Engineering common sense: you can design a simple system, reason about trade-offs, and know when not to over-engineer Self-directed: given a requirement and context, you can run with it and ask the right questions Useful background - Image data or signal processing (library is secondary — OpenCV, scikit-image, PIL, anything) - SQL and data warehousing (Snowflake, DBT, or similar) - CI/CD pipelines and packaging (GitHub Actions, uv, Docker) - Lightweight UI work in Python (NiceGUI, Streamlit, Gradio, or similar) - Experience in a regulated or quality-managed environment (medical devices, clinical, GxP, ISO, automotive, aerospace — any domain where software quality is formally managed) The archetype we’re hiring Your background might be biology, physics, engineering, or another quantitative field — you learned to code because you needed to, got good at it, and kept going. You’re more comfortable with a messy CSV and a domain problem than with a Kubernetes manifest. Classical CS background is absolutely fine too, but the domain orientation matters more than the degree