Dylan Caballero

Dylan Caballero

Full-stack engineer building production AI agents β€” grounded in enterprise identity and access-control integration.

I ship LLM-powered features that hold up in production β€” structured outputs, fallback chains, tool-calling orchestration β€” not demos that break on the second prompt.

Intro

  • πŸ“ Pembroke Pines, FL β€” South Florida (onsite capable)
  • πŸ’Ό Founder & Lead Engineer at Caballero Technologies
  • πŸ” 5 years integrating OAuth 2.0/OIDC & MFA against an enterprise IdP at Deloitte (Texas.gov)
  • πŸ€– Daily agentic dev tooling practice (Claude Code)
  • πŸŽ“ BA Computer Science, Florida International University

About

Full-stack engineer building production AI agents.

5+ years shipping production React and TypeScript on a high-availability platform used daily by millions of users. Today that experience goes into building AI agents and agent-driven workflows as core product work β€” not a side experiment, the main thing I ship.

I'm also a daily user of agentic development tooling β€” Claude Code and agent workflows are how I build β€” which gives me a working opinion on what actually accelerates engineers versus what looks impressive and gets abandoned.

My front-end work was built on hands-on authentication and access-control integration β€” implementing OAuth 2.0/OIDC and MFA flows against an enterprise identity provider (ForgeRock) β€” which maps directly onto the emerging problem of giving AI agents scoped, safe permissions.

AI agents & agent-driven workflows in productionOAuth 2.0/OIDC & MFA against an enterprise IdPDaily agentic tooling (Claude Code)

Skills

AI Agents & Generative AI

AI agent design & agent-driven workflowsMulti-agent orchestration (LangChain, LangGraph, CrewAI)OpenAI & Claude APIsPrompt engineeringTool/function callingStructured output validationFallback chains & retry logicState-aware multi-step LLM reasoningn8n workflow automation

Frontend

ReactNext.jsTypeScriptJavaScriptDesign-system-driven component developmentResponsive UIAccessibility (WCAG, ARIA)Client-side performance

Backend & APIs

Node.js/ExpressPython (Flask/FastAPI)REST API designWebhooksAsynchronous background workflowsPostgreSQLMicroservices

Developer Experience & Delivery

Agentic coding tools (Claude Code) in daily practiceGit/GitHubCI/CD (GitHub Actions)DockerLinuxCode reviewInternal tooling & developer workflow designAgile/Scrum

Security & Identity

OAuth 2.0OpenID ConnectJWTSSO/MFARBAC & permission scopingIdP integration (ForgeRock)Secrets handlingSecure API hardening

Selected Work

Production AI agents, client platforms, and full-stack builds.

Start a conversation

Babbydaddy.com

0β†’1 Β· Next.js / TypeScript / Python

A customer-facing AI product I built and launched 0β†’1: a guided React interface over state-aware, multi-step LLM reasoning that helps parents navigating custody, parenting time, and child support disputes prepare their case. I own authentication, data persistence, and deployment end to end.

Next.jsTypeScriptNode.jsPython

AI Resume Analyzer

Full-stack AI SaaS

Multi-agent LLM analysis pipelines with a React/Next.js front end, authentication, PostgreSQL persistence, and REST APIs. Autonomous agents collaborate to parse resumes, score them against job descriptions, and return actionable feedback β€” designed and shipped end to end.

LangChainLangGraphCrewAIReactNext.jsPostgreSQL

AI Voice Agent & Automation Platform

Production agent workflows

Production agent workflows orchestrating authenticated API calls across telephony, calendar, and CRM systems, with asynchronous background execution, retry handling, and fallback paths so pipelines behave predictably instead of failing silently. Private/client work β€” no public repo.

n8nWebhooksREST APIsAsync workflows

Cardboard Kings

Next.js / Postgres

A fan-focused sports card storefront where collectors can book cleaning services and sell prized collectibles.

Next.jsPostgresTailwind

Contact

Hiring for an AI Engineer, Forward Deployed Engineer, or Solutions Engineer role? Let’s talk.

I’m open to AI Engineer, Forward Deployed Engineer, and Solutions Engineer roles β€” especially where building production AI agents meets real identity and access-control problems.