Open to opportunities

Fabian Johnson

@fabianjohnson

Message

I build production LLM agents and scalable backend systems for live multi-tenant products.

Dominican Republic

Message

What I'm looking for

I'm looking for a senior remote AI Engineer role with a US-based company, building production LLM agents and RAG features with strong grounding, guardrails, observability, and safe multi-tenant data access.

I've shipped a production LLM agent at TOi that lets non-technical staff query live POS data—sales, tickets, and stock—in natural language across a platform handling roughly 16,000 orders per day.

I designed its tenant isolation so tenant identifiers never reach the model, are injected server-side from authenticated sessions, and make cross-tenant access structurally impossible. I also connected tools only to read replicas, keeping the assistant unable to mutate production data or compete with live order writes.

At GBH, I built FastAPI services and AI workflows with LangChain, LangGraph, and LangSmith for automation and data analysis, deploying serverless AI workloads on AWS and Azure. Earlier at Hatchworks, I built scalable NestJS and Rust services, React interfaces, microservice architectures, and mapping features using Google Maps, OpenStreetMap, and PostGIS.

Across nine-plus years, I've delivered APIs, integrations, cloud systems, and web applications using TypeScript, Python, Rust, AWS, Azure, MySQL, Redis, and DynamoDB. I'm focused on making LLM features grounded, observable, and safe for production use.

Experience

Work history, roles, and key accomplishments

Backend & AI Engineer

TOi

Dec 2024 - Sep 2025 (9 months)

Built and shipped a production LLM agent over a multi-tenant POS dataset using function/tool-calling, enabling natural language queries of live operational data. Designed a hard multi-tenant isolation boundary and wired all agent tools to a read-replica to prevent data mutation.