💼 Experience

• Lead the architecture and deployment of LLM-powered agentic systems for insurance claims automation, utilizing LangChain, LangGraph, and Streamlit to orchestrate multi-agent workflows and user-facing prototypes.
• Designed and implemented Python/FastAPI-based microservices supporting stateful, event-driven Q&A workflows modeled via ClaimState (Pydantic), enabling agents to persist, reference, and evolve structured claim context across interactions.
• Built TrussController, a modular middleware layer that enforces compliance through configurable PII redaction, role-based access controls, agent-aware routing, and LLM provider abstraction—powering seamless integration with OpenAI, Anthropic, and open-source models.
• Conduct LLM performance evaluations and benchmarking via prompt engineering
• Conducted systematic LLM evaluations, fine-tuning prompt templates with chain-of-thought reasoning, handling ambiguity, and optimizing completions for correctness, latency, and cost—resulting in measurable accuracy improvements in claims triage and policy recommendations.
• Delivered scalable, observable, and secure LLM infrastructure with real-time tracing, audit trails, fallback strategies, and interactive dashboards for redaction decisions and agent-level cost breakdowns—supporting enterprise-grade deployment standards.
• Prototyped graph-based memory systems (e.g., temporal graphs for inter-agent knowledge transfer) and integrated vector databases to persist memory, context embeddings, and prior case knowledge for longitudinal reasoning.

Transformed a monolithic defense software platform into a high-compliance, cloud-native SaaS solution serving elite clients (Pentagon, White House), while driving innovation, scalability, and security.

• Led technology strategy for healthcare/fintech platform used by Miami Cancer Institute
• Developed multi-tenant cloud architecture with Stripe PCI compliance
• Built HIPAA-compliant AI systems for patient engagement
• Improved agile velocity by 30% and led investor due diligence

• Led cloud migration of legacy defense systems to modern architecture
• Introduced DevOps practices including CI/CD and infrastructure as code
• Built predictive analytics applications for national security
• Mentored high-performing engineers through code reviews and pairing

• Developed security clearance and compliance products
• Modernized systems using .NET and JavaScript stacks
• Created automation tools to boost developer productivity

• Built billing and provisioning systems with fintech infrastructure
• Developed robust backend services and database layers