Projects, Tools, and Impact
I work on applied projects that translate machine learning into operational systems used by governments and organizations.
These projects are designed not just to build models, but to support decisions and improve outcomes.
Homelessness Prevention (Allegheny County)
- Built ML system to identify households at risk
- Integrated into service delivery workflows
- Currently being evaluated through a randomized controlled trial
Behavioral Health Outreach (Kansas)
- Identified individuals at risk of crisis or system involvement
- Supported proactive outreach and intervention
- Reduced unnecessary jail bookings
Lead Exposure Prevention (Chicago)
- Predicted homes at highest risk of lead exposure
- Enabled targeted inspections
- Demonstrated improved detection rates
Tools and Systems Developed
- Risk stratification models for public systems
- Case prioritization tools for frontline workers
- Policy simulation and evaluation frameworks
What Makes These Work
- Close collaboration with domain experts
- Integration into real workflows
- Continuous evaluation and iteration
- Focus on decisions, not just models