Case study · Healthcare
Claim-denial prediction
A model that flags insurance claims likely to be denied before they are submitted, served to internal apps through a monitored API.
- Client
- US healthcare revenue-cycle group
- Result
- Ran in production with monitoring
- Stack
- Python, SQL, scikit-learn, REST APIs, Dashboards
The problem
Denied claims cost the revenue-cycle team time and cash. They needed to know which claims were at risk, and why, before submission and inside the tools they already used.
The work
- ETL and feature pipelines in SQL and Python over the claims data.
- Classification models that score denial risk for each claim.
- A REST API so internal applications could call the model directly.
- Production monitoring and issue triage, plus latency and reliability fixes with the engineering team.
- Real-time dashboards for operations and leadership: claims flow, denial drivers and turnaround time.
Outcome
The model ran in production under monitoring, and operations and leadership saw the main denial drivers every day.
The client isn’t named, to respect confidentiality. I can share more detail on a call where permitted.
Talk about a similar projectMore work
Support assistant in Microsoft Teams
A Teams and telephony assistant that answers support questions from the company’s own documents, using an OCR and entity-extraction pipeline underneath.
Live in a corporate Teams tenant
AI due diligence for investors
Technical reviews of early-stage AI startups for venture investors, plus retrieval tools that turn a data room into a first-draft investment memo.
Seed and Series A deals