remote | formerly Thoughtful AI | healthcare RCM · agentic AI platform
AI-powered revenue cycle management platform combining agentic AI, human-in-the-loop agents, and clinical ontology to automate healthcare administrative and financial workflows.
- Manage 5 concurrent AI agent delivery programs across 4 enterprise healthcare clients, spanning 3 agent types and 4 EHR integration platforms, with cross-functional teams of 6-12 per project (engineering, QA, product, customer success, and customer-side billing/IT/clinical stakeholders).
- Delivered infrastructure capacity analysis covering baseline performance, 15x scaling projections, and API call volume modeling across 4 Azure VM pricing tiers, enabling leadership to right-size infrastructure costs from an initial $5.1-8.3M estimate down to $3.2M/month.
- Authored the technical business case to replace two fragile, coupled production agents with a unified API-first architecture and two reusable services, quantifying the gap (2,300+ lines of brittle desktop automation vs. 200ms API calls) and designing a 3-track implementation roadmap that secured stakeholder buy-in for a multi-month development investment.
- Led UAT coordination across 10 regional testers (Texas, Upper Midwest, Central), creating region-specific Notion testing documentation, triaging 25+ feedback items per cycle, and converting findings into Jira tickets with structured acceptance criteria.
- Led VOB + CPT Calculator integration discovery sessions, identifying API endpoints and HST database access requirements, bridging customer IT constraints with engineering implementation needs for a net-new authorization workflow.
- Identified 1,263 unprocessed commercial appointments through appointment data analysis with the operations team, removing a Medicare-only filter that had constrained processing scope and shifting the payer mix strategy for the following month's automation pipeline.
- Coached team members on stakeholder readiness: preparing hard numbers, cost comparisons, and timeline projections before presenting to client leadership rather than improvising in live calls.
- Invented a modular requirements methodology (ARD) replacing traditional PRDs company-wide, combining visual process flows with structured databases, hierarchical decomposition, and 19 automated QA validation checks. Adopted across all 5 active projects as the company standard.
- Developed 27 GenAI-powered workflow automations (Claude Code) covering daily operational pipelines, weekly reporting with Notion sync, change impact analysis, and Jira ticket generation. Built 3 production Python systems for analytics dashboards, business rules validation, and deterministic workflow execution.