Every ML Engineer Has a Portfolio.
Almost None of Them Include Healthcare.
The Clinical AI Portfolio Accelerator. 8 weeks. A real clinician partner. A deployed app that gets you into health tech.
The Problem
Healthcare AI is one of the fastest-growing and highest-paying niches in machine learning. But breaking in is hard — not because the ML is harder, but because you don't have access to the clinical context that makes the work meaningful.
You can train a model on any dataset. But can you build one that a physician would actually trust? That's the gap between a Kaggle project and a healthcare AI career.
Hiring managers in health tech don't want another MNIST project. They want proof you can work with clinical data, clinical workflows, and clinical stakeholders.
What This Is
The Clinical AI Portfolio Accelerator is an 8-week program where you — an engineer — partner with a real clinician to build a deployed clinical AI system together.
Your clinician partner brings domain expertise: what the data means, which predictions matter, whether the model's output makes clinical sense. You bring the engineering: data pipelines, model architecture, evaluation, deployment.
No healthcare experience required. That's the whole point — you learn it by building alongside someone who lives it.
Your Role as Technical Lead
- Architect the data pipeline — ingest, clean, and transform clinical datasets into model-ready features
- Build and train the model — select architecture, tune hyperparameters, iterate on clinical feedback
- Implement evaluation beyond accuracy — precision, recall, calibration, and clinical relevance metrics
- Deploy the application — build a working prototype with a live URL
- Co-author a publishable case study documenting the technical and clinical decisions
What You Walk Away With
Your deliverables
- A deployed healthcare AI application with a live URL — not a notebook, a working system
- A GitHub repository with end-to-end clinical ML: data processing, training, evaluation, deployment
- A publishable case study co-authored with a clinician — proof of cross-functional capability
- Real clinical domain knowledge — EHR data structures, clinical workflows, regulatory constraints
- A differentiated portfolio that separates you from every other ML engineer in health tech hiring
- Deep applied ML foundations — you'll build and deploy a real predictive AI pipeline end to end: data preparation, feature engineering, model training, clinical evaluation, and deployment. Not a tutorial. A working system in production.
Choose Your Clinical Use Case
Each clinician-engineer pair selects a real clinical prediction problem:
These are sample use cases. If your clinician partner has a prediction problem from their own specialty, you'll build that instead — the program supports any well-scoped predictive AI use case grounded in clinical data.
Tools & Stack
8-Week Structure
| Weeks | Focus |
|---|---|
| 1–2 | Scope the problem with your clinician. Set up data pipelines. Explore the clinical dataset. |
| 3–4 | Feature engineering. Train initial models. Get clinical feedback on what the outputs mean. |
| 5–6 | Model refinement. Evaluation beyond accuracy — clinical relevance, calibration, error analysis. |
| 7 | Deploy the application. Build the live prototype. Prepare the case study and GitHub repo. |
| 8 | Demo Day — present your system to the cohort. Finalize documentation and publishable case study. |
Time commitment: 4–5 hours/week — 2 hrs live session (weekends) + 2–3 hrs async with your clinician partner.
Who This Is For
- ML/data engineers who want to break into healthcare AI with a portfolio project that proves domain competence
- Software engineers with Python and basic ML knowledge who want to specialize in one of the highest-demand verticals
- Any technical builder tired of generic portfolio projects and ready to work on something that matters
Who's Behind This
1:1 clinician-engineer pairing
Starts June 15, 2026
Application is brief. We select engineers committed to shipping — not just learning.
Questions? roofi@gluonlabs.ai