The Role

• You'll build the vision models that power 5C's autonomous radiology pipeline — from detection and segmentation on CT/MRI to classification on X-ray and mammography. Every model you ship goes directly into clinical workflows where radiologists and patients depend on it.

• This is applied science with real stakes. You'll own problems end-to-end: from dataset curation and architecture design through training, evaluation, and production deployment.

What You'll Own

• Design, train, and deploy vision models for medical image analysis across multiple modalities

• Build multi-modal systems that combine medical images with clinical text and structured metadata

• Develop evaluation pipelines that measure clinical accuracy, not just benchmark metrics

• Collaborate with radiologists to define ground truth, validate outputs, and iterate on model behavior

• Contribute to research publications and technical blog posts

• Mentor junior scientists and raise the bar for model quality across the team

You Should Have

• 4+ years of applied research or engineering in computer vision

• Strong PyTorch expertise — custom architectures, distributed training, ONNX export

• Experience with at least one: volumetric (3D) analysis, vision transformers, medical imaging pipelines (DICOM, NIfTI)

• Published research or production models in computer vision or medical imaging

• Demonstrated ability to take a model from research prototype to production

Why 5C

• 5C is one of the best places to work for serious, engaged builders. We are a leading company in agent-driven software development and a leading business in AI for medical imaging — both at the same time

• Small, highly empowered teams. High execution capability, high taste, high agency. No committees, no approval chains — just builders who ship

• Resumes are not important. Proof of work and proof of a high learning rate is everything we look for. Show us what you've built

• Your models will be used by thousands of radiologists across India — directly impacting patient outcomes