About Amplifier AI
We're building the next generation of surgical planning software for interventional radiology and cardiothoracic surgery. Our platform turns CT scans into interactive 3D models that help surgeons plan complex procedures — from endovascular aortic repair (EVAR) to transcatheter aortic valve implantation (TAVI) and beyond.
Our core library powers 17+ clinical analysis pipelines that perform automated anatomical segmentation, vessel centerline extraction, hemodynamic measurements, and risk stratification. Every algorithm we ship has a direct impact on patient outcomes.
We're a small, focused team where your code goes from PR to operating-room screen in weeks, not quarters. If you want to work on technically challenging problems that genuinely matter, keep reading.
What You'll Work On
This isn't a generic backend role. Here's what a typical month might look like:
Extending our NNUnet-based segmentation pipelines (we run 15+ trained models across organs, vessels, bones, and tumors) and validating results against clinical ground truth
Building automated measurement workflows — think aortic neck angulation, iliac tortuosity indices, or landing-zone diameter calculations that feed directly into surgical reports
Developing new workspace modules: designing the data flow from raw DICOM volumes through preprocessing, AI inference, post-processing, and 3D scene generation
Optimizing inference performance for GPU and CPU execution, including memory-aware scheduling and adaptive resolution strategies
Working with SimpleITK, NumPy, and our in-house volume processing library to manipulate medical images at the voxel level — resampling, label morphology, island filtering, ROI extraction
Collaborating directly with radiologists and surgeons to translate clinical requirements into algorithmic specifications
Our Tech Stack
Language: Python 3.12+ (modern typing, dataclasses, async patterns)
Medical Imaging: SimpleITK, NIfTI volumes, DICOM processing
AI/ML: NNUnet framework, PyTorch, custom inference pipelines
Infrastructure: Docker, NVIDIA GPU compute, CI/CD with automated testing
3D Visualization: Unity integration (you won't write C#, but your Python outputs drive the 3D viewer)
Tools: Git, Jira, Confluence, VS Code / PyCharm
What We're Looking For
Must-haves:
Strong Python fundamentals: you write clean, typed, well-structured code and understand performance trade-offs
Experience with NumPy/SciPy for numerical computing — you're comfortable thinking in arrays and transforms
Familiarity with medical imaging concepts (DICOM, NIfTI, voxel spaces, coordinate systems) or strong willingness to learn quickly
Understanding of machine learning inference pipelines, even if you're not training models from scratch
Ability to work independently in a remote-first environment with 4+ hours of overlap with CET business hours
Nice-to-haves:
Experience with SimpleITK or ITK for image registration, segmentation, or morphological operations
Background in NNUnet or similar medical segmentation frameworks
Knowledge of 3D geometry: meshes, centerline extraction, spatial transforms
Contributions to open-source scientific Python projects
Domain experience in radiology, surgical planning, or health-tech
A Few Things to Know
The codebase is scientifically dense — you'll work with Hounsfield units, anatomical coordinate systems, and clinical measurement standards. Curiosity about the domain is essential.
We're a startup. Processes are evolving, requirements shift, and you'll sometimes need to figure things out from first principles rather than following established playbooks.
This is primarily a backend/algorithmic role. If you're looking for frontend or web development work, this isn't the right fit.
We care deeply about code quality — type hints, meaningful tests, clear documentation. We review each other's work carefully.
Why Join Us
Direct clinical impact: your code helps surgeons plan life-saving procedures more accurately
Technically deep work at the intersection of AI, medical imaging, and 3D visualization
Small team, big ownership: you'll shape architecture decisions and see your ideas go to production
Flexible remote setup with asynchronous communication and minimal meetings
Competitive compensation with equity participation
How to Apply
Send us your CV along with a short note covering:
A brief description of the most technically challenging Python project you've worked on — what made it hard, and how you solved it.
Your experience (if any) with medical imaging, numerical computing, or ML inference — even if it's from a different domain.
Your availability and time zone.
No cover letter template needed — just be direct. We review every application personally.