I'm Andrew.

As the Lead AI/ML Engineer at General Atomics, I lead the enterprise AI program and serve as the organization's subject-matter expert on LLMs — designing, building, and shipping secure generative-AI platforms that transform company operations.

• Avoided $15M/yr in spend by developing an in-house, DoD-compliant enterprise AI chatbot using AWS Bedrock, Azure AI Foundry, and LiteLLM.
• Achieved 4x workflow efficiency for 10,000+ users with a self-service RAG platform built on Azure AI Search, AWS RDS PostgreSQL, and Python.
• Delivered $50M ROI with a self-service agent platform using MCP and a centralized authentication/authorization framework integrated with internal systems.
• Cut technical-order development time by 40% with agentic search and generation powered by GraphRAG on Amazon OpenSearch, Amazon Neptune, and AWS Bedrock.
• Reduced deployment time by 90% through infrastructure-as-code with Terraform/Terragrunt and GitHub Actions CI/CD.
• Eliminated 90% of manual processing across key workflows with autonomous multi-agent systems built on Azure Durable Functions, Python, LangGraph, Semantic Kernel, and the Claude SDK.
• Boosted developer productivity by 30% by rolling out AI development tools and integrating them into the SDLC while maintaining security and trustworthiness.

Working as a Systems Administrator, I was tasked with improving and modernizing the company's IT infrastructure — bringing concepts of DevOps and Agile development to automate, innovate, and quickly generate value for over 15,000 end users.

• Spearheaded the team that implemented generative AI across the enterprise — the groundwork for the company's AI agent program.
• Personally developed generative AI-powered applications that improved access to information for technicians and executives.

As an IT strategic analyst, it was my job to be tier 1 helpdesk, systems administrator, and automation engineer all at once — anticipating the needs of end users, responding to them, and implementing durable solutions.

• Reduced onboarding and offboarding time by 90% by automating account provisioning across all company applications.
• Implemented a centralized MuleSoft API management system, improving API discoverability, security, and auditability.
• Developed internal tools using Python and JavaScript that significantly improved team efficiency.

At Reynolds & Reynolds, our clients were car dealerships across the country. I was tasked with auditing their compliance with all relevant authorities and advising them on marketing strategy.

• Managed a territory of 100+ clients generating $1M+ in annual sales.
• Performed compliance audits across local, state, federal, and corporate authorities — resolving failures quickly and advising clients on preventing future violations.
• Supported client marketing strategy and content creation alongside the compliance practice.

Machine learning and computing systems specialization.

Coursework in research methods, statistics, and cognition.