Leo Morgan

Led frontend architecture using React to build scalable, responsive enterprise UI interfaces, applying advanced state management patterns (Redux Toolkit, Context API), performance profiling, memoization, lazy loading, and reusable design-system components. Developed backend RESTful APIs with Python (FastAPI), including validation, authentication (JWT/OAuth), and database transactions with PostgreSQL schema design and query optimization. Debugged and refactored legacy Python services and frontend modules to reduce production incidents. Built production RAG pipelines using LangChain and LangGraph, integrating vector databases (FAISS/Pinecone) and implementing deterministic agent orchestration via LangGraph state machines with retries and fallback handling. Deployed and maintained end-to-end AI workflows supported by robust reliability and scalability practices.

Built LLM-powered intelligent agents for customer support and internal operations, reducing manual workload (~60%) and improving response times. Designed agent pipelines for intent detection, decision logic, tool execution, and response generation, improving automation reliability and consistency. Implemented RAG-enabled agent systems that retrieved enterprise knowledge dynamically to improve answer accuracy and reduce hallucinations. Created robust Python orchestration frameworks for agent state management, retries, fallbacks, and execution paths to increase resilience. Deployed and monitored agentic AI services on AWS using Docker and Kubernetes to achieve high availability and scalable throughput, collaborating with product, data, and platform teams to integrate agent-based AI into enterprise SaaS workflows.

Developed scalable backend services using Django and Flask for data-intensive applications and ML-driven features. Improved database schema design and query performance to reduce system response times and database load (~30%). Implemented CI/CD pipelines and automated testing to increase release reliability and reduce production defects. Built and deployed secure, auditable LLM-powered agent systems for automating and validating complex internal financial documentation workflows, supporting data consistency and internal compliance. Delivered modern React UIs for internal tools and analytics, including routing, controlled forms, API integration, and frontend validation/error handling. Refactored legacy Angular modules into React-based architecture to improve maintainability and performance (~30%). Integrated event-driven pipelines using Kafka for scalability and decoupling.