Open to opportunities

Taylor Williams

@taylorwilliams5

I build production AI platforms and high-throughput backend systems.

I've built production AI and backend platforms at Starbucks, Optum, Hugging Face, Stripe, and Uber across healthcare, financial transactions, operational intelligence, and large-scale telemetry.

At Starbucks, I architect Python and Java microservices for LLM orchestration and real-time operational decisions, connecting LangChain, LlamaIndex, PyTorch, Hugging Face, Kafka, vector databases, and Kubernetes across AWS and Azure.

At Optum, I led CareScribe backend architecture for a HIPAA-compliant ambient clinical intelligence platform, building distributed LLM serving with Ray Serve, vLLM, GPU inference, OpenAI-compatible streaming APIs, and EHR integrations.

I've also delivered scalable AI infrastructure at Hugging Face and resilient transaction and telemetry systems at Stripe and Uber, with a focus on fault tolerance, observability, distributed systems, and high-throughput production workloads.

Experience

Work history, roles, and key accomplishments

Current

Senior Software Engineer - AI & Backend Architecture

Current

Starbucks

Mar 2025 - Present (1 year 6 months)

Architected Python and Java microservices hosting LLM orchestration pipelines and real-time operational decision systems. Optimized AI model deployment using Triton Inference Server and Docker on Kubernetes, and designed Kafka-based event-driven architectures for RAG.