AI Diagnostic Engine
AI Diagnostic Engine: Build a Health Prediction Model for a HealthTech Startup. A 4–8 weeks Intravarsity career experience (30–50 hours) for school and underg
You'll be matched with a healthtech startup from our San Francisco campus ecosystem. Using their anonymized patient datasets, you'll build classification models for early disease detection — the same type of work their data science team does. Your model and analysis become part of the startup's research pipeline.
Work remotely with a Silicon Valley healthtech startup to build a machine learning model that predicts disease risk from real anonymized patient data. Not a course — a real R&D project.
Who it is for
High school, secondary school students & undergraduates. No coding experience required — we teach everything from scratch.
Weekly plan
Week 1: Startup Kickoff & Data Exploration
- Meet your healthtech startup team and understand their clinical challenge
- Set up Python environment and explore the anonymized health dataset
- Clean data with Pandas, identify missing values and correlations
- Visualize health risk factors with Matplotlib and Seaborn
Week 2: Model Development
- Build logistic regression classifiers for disease prediction
- Train decision tree, random forest, and XGBoost models
- Evaluate with accuracy, precision, recall, F1, and ROC-AUC
- Cross-validate and handle class imbalance with the data team
Week 3: Deployment & Dashboard
- Build a Streamlit dashboard for the prediction model
- Add interactive visualizations and patient risk scoring
- Workshop with startup CTO on production deployment considerations
- Deep dive into AI ethics — bias, privacy, fairness in healthcare
Week 4: Presentation to Startup Leadership
- Document methodology in a technical report
- Present findings to the startup's founders and medical advisors
- Discuss real-world deployment considerations and next steps
- Receive Proof-of-Work credential and startup recommendation
What you produce
- ML prediction model with documented methodology
- Interactive Streamlit dashboard
- Technical analysis report
- Startup recommendation letter
- Intravarsity Proof-of-Work credential
Skills
Python, Pandas, Scikit-learn, XGBoost, Streamlit, Healthcare Analytics, Data Visualization, Ethical AI
Commitment
~5–8 hours/week for 4–8 weeks (30–50 hours total). 100% Virtual — work remotely with a Silicon Valley startup.