hello, I'm
Rena Thomas
Senior Full-Stack Engineer
AI Systems & LLM Infrastructure · Platform Engineering & DevOps
Over a decade building and scaling production systems across fintech, healthcare, and AI infrastructure. Founding CTO turned LLM specialist — I build things that are observable, resilient, and designed to compound.
About
I'm a full-stack engineer based in New York with over a decade of production-grade experience building systems that scale. My background spans fintech at JP Morgan, five years as founding CTO at Vegannation, and now deep specialization in LLM agent architecture and real-time AI orchestration as a senior contractor.
I got into software through a physics and mathematics degree at Yeshiva University and a masters in computer science from Ben Gurion University. The rigor from that path shapes how I approach every system I build — I care deeply about correctness, observability, and designing for failure.
Right now I'm most interested in the infrastructure layer underneath AI products: agent orchestration, multi-provider voice AI, RAG pipelines, and the reliability engineering that makes AI systems actually trustworthy in production.
Location
New York, NY
Currently
Freelance / Contractor
LLM systems, AI infra
Education
MA Computer Science
Ben Gurion University
BA Math & Physics
Yeshiva University
Tech Stack
AI & LLM
- LLM Agent Orchestration
- Multi-Agent Pipelines
- RAG Architecture
- OpenAI · Whisper
- Deepgram · Voice AI
- Prompt Engineering
- TensorFlow · PyTorch
Cloud & DevOps
- AWS (EKS, Lambda, API GW)
- GCP · Azure · IBM Cloud
- Terraform · IaC
- Docker · Kubernetes
- CI/CD · GitHub Actions
- Observability & SRE
Languages
- TypeScript · JavaScript
- Node.js · NestJS
- Python (FastAPI, Flask)
- Java (Spring Boot)
- Solidity · Perl
Frontend & Data
- Next.js · React · Vue
- React Native · Angular
- Tailwind CSS
- PostgreSQL · Firebase
- Real-Time Sync
Recent Posts
All posts →Fable 5 vs Opus 5: Picking a Frontier Model for Real Engineering Work
Anthropic shipped two frontier-tier models seven weeks apart, and Opus 5 lands at half of Fable 5's price. The interesting question isn't which is smarter — it's where the premium actually pays for itself.
Prompt Injection Is a Systems Problem
Two years of trying to solve prompt injection at the prompt layer has produced no reliable defense. The teams shipping agents safely stopped trying — and constrained the blast radius instead.
Context Engineering Beats Prompt Engineering
Million-token context windows didn't kill retrieval. They moved the problem: from 'what can we fit' to 'what should be in there', which is a much harder question.