Greg Thomas

I’m a Systems Architect and Engineer focused on large language model (LLM) orchestration, custom Domain-Specific Languages (DSLs), and high-performance computing pipelines. I bridge legacy hardware constraints with modern distributed AI infrastructure, designing highly optimized asynchronous runtimes and deploying agentic workflows on Google Kubernetes Engine (GKE).

In my recent work, I architected a high-performance compute pipeline that integrates ChorusDSL with a Fortran 90 and x64 translation layer to enable fast Fourier transformations on restrictive CPU bus architectures—while balancing deterministic math compute with stochastic model execution. I also designed an 8-way asynchronous execution engine for autonomous systems, then extended runtime capabilities into cloud environments using Kubernetes and CI/CD pipelines.

Earlier, as Lead Developer & Digital Architect at Mandelbroccoli Artworks, I transitioned core studio technology from geometric art generation into advanced Generative AI systems, including VQGAN+CLIP visual production pipelines and a high-availability 2TB SMB storage pool across 10 hardware nodes. I complement engineering with digital content and technical writing—building automated multimedia workflows and producing deeply researched, SEO-optimized documentation via WordPress.