Alexander Weaver

I’m a senior full-stack and platform engineer focused on AI-driven distributed systems, real-time runtimes, and high-performance backend architecture. I thrive on building zero-to-one platforms that balance performance, cost, and reliability while staying security- and compliance-minded.

At MaxHome.AI, I engineered the architecture, infrastructure, and execution environment for an AI automation platform—building backend, runtime, and agent-orchestration layers with event-driven pipelines and encryption. I designed a secure, multi-layer sandbox for executing LLM-generated Python and TypeScript, using V8 isolates for ~20 ms low-latency execution and Firecracker microVMs for ~2 s cold-boot isolation. I also shipped real-time conversational experiences with WebRTC and built a multimodal STT → LLM → TTS pipeline using Gemini Live API, OpenAI Realtime, and ElevenLabs.

Previously at OPTUM, I built LLM-powered helper tooling that supported operational diagnosis across patient events, payments, CRM records, orders, and distributed system failures. I implemented retrieval-grounded workflows with pgvector-based semantic retrieval, added AI observability (drift detection, latency monitoring, structured model-behavior logging), and delivered HIPAA-compliant backend services with RBAC, encryption, and audit logging to keep PHI flows safe for regulated environments.

Earlier, at MarqVision, I designed high-performance storefronts and merchant dashboards, implemented backend microservices (Node.js, Go, Python/FastAPI), and architected real-time inventory and pricing with event-driven systems. I improved checkout performance by 35%, helped maintain 99.9% uptime during peak retail events, and carried that same performance/testing mindset from my Google work through tools like Jest and Cypress—always aiming for production-grade reliability.