Full Stack Engineer – Medical Imaging Web Platform & 3D Viewer

We are building a cloud-native, AI-assisted radiology platform for CT and MRI. The product has two equally important components: a browser-based 3D viewer (Cornerstone3D / VTK.js) and the web application surrounding it, covering worklists, study management, clinical workflows, user and access management, and an agentic AI assistant.

The product is being developed as a regulated medical device, with a target of FDA 510(k) clearance.

As a Full Stack Engineer on the imaging team, you will split your time between the application and the viewer. You will build the web application and its backend services, and work on the 3D viewer where AI results and clinician workflows come together.

You will collaborate closely with radiologists, ML engineers, product, QA, and our US-based engineering team.

1. → Web Application and User Workflows

• →Build and own the React/TypeScript web application, including study worklists, search and filtering, study and series management, case assignment, review and approval queues, and reporting screens.
• →Design and implement clean, fast, accessible UIs for clinical users, including the interface for the agentic AI assistant, with chat, suggested actions, and review of proposed actions.
• →Implement authentication and role-based access control (OIDC/SSO, Keycloak, or similar), user and tenant administration, and audit-log views.
• →Build real-time and long-running workflow experiences, including job status, progress indicators, notifications, and retries for processing and AI pipelines.
• →Manage application state, API integration, and caching to keep large worklists and studies responsive.
• →Create annotation, labeling, and quality-review tools that capture clinician feedback in a structured format.

1. → Viewer and Rendering

• →Extend and maintain the web-based 3D viewer (Cornerstone3D / VTK.js) with DICOMweb integration (QIDO-RS, WADO-RS, STOW-RS).
• →Implement volume rendering, surface rendering, and MIP/MinIP for CT and MRI, including transfer function presets and controls.
• →Build multiplanar reconstruction (MPR), oblique, curved, and straight reformats with synchronized crosshairs and cross-reference navigation.
• →Develop interactive tools for measurements, annotations, clipping, cropping, and segmentation editing, including brush, threshold, and region-growing tools.
• →Optimize viewer performance for large, multi-series studies through progressive and streaming loading, GPU memory budgeting, web workers, and caching.

1. → Advanced Clinical Workflows

• →Build vessel segmentation and reconstruction workflows, including centerline extraction and curved planar reformation (CPR) views.
• →Develop procedure-planning features such as landmark placement, device sizing, and trajectory planning.

1. → AI Integration

• →Integrate the viewer with AI services for segmentation, detection, and the agentic AI assistant through secure, well-defined APIs.
• →Read and write AI outputs as standard DICOM objects, including DICOM-SEG, Structured Reports such as TID 1500, and RT Structure Sets, and display them as editable overlays.
• →Build human-in-the-loop workflows where clinicians review, edit, accept, or reject AI results, with provenance tracking and an audit trail of every change.
• →Ensure AI-generated content is clearly distinguishable from clinician-entered content in the UI.

1. → Backend and Platform

• →Design and build backend services and REST APIs using Python/FastAPI or Node.js for study management, workflow orchestration, users, and AI results.
• →Model and manage data in PostgreSQL and object storage (S3), and build pipelines for ingesting, indexing, and serving large imaging datasets.
• →Integrate with PACS/VNA and DICOM sources using DICOMweb, Orthanc, or similar technologies, and with workflow engines such as Temporal.
• →Deploy and operate services in containers on AWS and Kubernetes alongside the platform team, with logging, monitoring, and CI/CD.

1. → Quality and Compliance

• →Develop software under IEC 62304, including traceable requirements, unit, integration and end-to-end tests, and code reviews.
• →Contribute design and verification documentation to support regulatory submissions.
• →Handle protected health information (PHI) securely and follow HIPAA-aligned engineering practices.

• 4+ years of professional full stack development experience, with a track record of shipping production web applications end to end.
• Strong TypeScript and React skills, including state management, component design, and browser performance optimization involving rendering, memory, and web workers.
• Solid backend experience with Python (FastAPI preferred) or Node.js, REST API design, authentication and authorization (OAuth2/OIDC, RBAC), and PostgreSQL.
• Experience with Docker, Git, CI/CD, and automated testing, including unit, integration, and end-to-end testing.
• Hands-on experience with Cornerstone3D/Cornerstone.js, OHIF, VTK.js, or comparable imaging stacks such as ITK, 3D Slicer, or WebGL volume rendering. At least 1–2 years in medical imaging or 3D visualization is a strong plus.
• Working knowledge of DICOM, including the data model, key tags, and patient coordinate systems, as well as DICOMweb and PACS.
• Comfort with 3D concepts, including transforms, meshes, volume rendering, and MPR geometry.
• Experience consuming ML/AI model outputs in a production application.

• Experience with vessel analysis, centerline/CPR, or surgical/interventional planning software.
• Contributions to open-source imaging projects such as Cornerstone3D, OHIF, VTK.js, dcmjs, or similar.
• WebGL/WebGPU and GLSL/WGSL shader programming experience.
• Familiarity with MONAI, nnU-Net, TotalSegmentator, or MedSAM, including prompt-based and interactive segmentation.
• Experience with DICOM tooling such as Orthanc, dcm4che, pydicom, highdicom, or dcmjs.
• Background in regulated environments, including IEC 62304, ISO 13485, FDA 510(k), CE-MDR, and HIPAA/GDPR, with an understanding of requirements for AI-enabled medical devices.
• Familiarity with HL7 FHIR, Kubernetes, AWS, Keycloak, and workflow engines such as Temporal.
• Experience building LLM- or agent-based user experiences, including chat UIs, tool-use review, and streaming responses.