AI Full Stack Engineer | Python · FastAPI · Next.js · OpenAI

From manual process to the system that runs it.

I worked remotely from Mexico for Handbook, a US company, where I developed features for an AI workflow automation platform and worked directly with customers: gathering requirements, turning them into tasks and shipping them with FastAPI, PostgreSQL, Next.js, TypeScript and OpenAI Agents.

See real-world experience Get in touch

Available immediately · Mexico City · Remote (CST) · AI Full Stack / AI Systems / Product Engineer

OpenAI Agents · Function calling · Structured outputsFastAPI · Next.js · PostgreSQLPytest · Jest · Playwright

Latest work

Handbook, Feb–Jul 2026.

An AI workflow automation platform for operations with handoffs, evidence and SLAs. What I built there:

Cyclic workflows: iteration guards, recovery, versioned snapshots and execution history
Conversational LLM intake plus audio transcription to generate workflows
Visual graph editor (React Flow + ELK.js) with real-time collaboration and autosave

Latest role

AI Full-Stack Engineer at Handbook

Feb–Jul 2026 · Remote from Mexico.

Main stack

Python and TypeScript

FastAPI and PostgreSQL on the backend; Next.js and React on the frontend.

Experience

3+ years

From customer requirements to production: product, data and AI.

Who I am

From requirements to production.

I have spent three years building software that runs businesses: LLM-automated support, data pipelines, workflow platforms and a digital election. Before writing code, I talk to the people who operate the process.

In short

I talk to the customer, model the process, build the system.

Product with Next.js and TypeScript; backend with FastAPI or Nest.js on PostgreSQL; data orchestration with Airflow; and LLMs with agents, structured outputs and human-in-the-loop.

FastAPI + Pydantic + SQLAlchemy and Nest.js, on PostgreSQL
OpenAI Agents with structured outputs and human-in-the-loop
Tests and CI as a baseline: Playwright, Pytest, Jest, GitHub Actions

Latest role

AI Full-Stack Engineer at Handbook (AIPlaybookHQ), Feb–Jul 2026

Based in

Mexico City · remote or hybrid

Specialty

LLM workflows: agents, function calling, structured outputs and human-in-the-loop

Looking for

AI Full Stack, AI Systems or Product Engineer — available immediately

I gather requirements with the customer

I talk directly with customers and operators: what they do, where it gets stuck, what evidence they need. That turns into technical tasks and workflows to automate.

LLMs in production

I integrate LLMs into operational flows with function calling, structured outputs, traceability and human intervention points.

Data and quality as a baseline

Airflow DAGs for analytics, reliable ETL and a real testing culture: Playwright, Pytest, Jest, React Testing Library, MSW, Pyright and Ruff.

Real cases

Four systems in production, across different industries.

LLM workflows, automated support, analytics pipelines and a digital election: each case with context, my exact role and what was left running.

handbookai.io · Abrir

handbookai.io

Website

handbookai.io

Guided execution platform for complex processes with handoffs, evidence, SLAs and traceability.

Active case

LLM workflow automation at Handbook

A platform that turns processes with handoffs, evidence and SLAs into guided execution. I developed workflow builder features and coordinated delivery with customers — FastAPI, PostgreSQL, Next.js, React and OpenAI Agents.

FastAPIOpenAI AgentsNext.jsPostgreSQL

Role

AI Full-Stack Engineer · Product Architecture & LLM Systems

Company / context

Handbook / AIPlaybookHQ

Period

Feb 2026 - Jul 2026

Problem

Manufacturing, banking and insurance operations live in emails and spreadsheets: missed steps, untracked SLAs and zero traceability of who did what.

My role

Full stack development and customer coordination: I gathered and clarified requirements, turned them into prioritized tasks and coordinated delivery.

What I built

Cyclic workflows with guards and snapshots, LLM intake with audio transcription, dynamic forms on case schemas, configurable HTTP integrations, email templates (Gmail, Outlook, Zoho) and real-time collaboration.

What was left running

Workflow execution with cycles: iteration guards, recovery, versioned snapshots and execution history.
Conversational LLM intake plus audio ingestion/transcription to turn unstructured sources into executable workflows.
Visual graph editor (React Flow + ELK.js) with real-time collaboration, autosave and drag/serialization performance work.
Dynamic forms on case schemas and configurable HTTP integrations with request/response mappings.
Backend and frontend tests with Pytest, Jest, React Testing Library, MSW and Playwright.

Personal projects

Technical demos you can try.

Experiments built to practice specific concepts. Every project includes a working demo and its source code.

concurrency / Web Workers

Concurrency applied to a web game

A web game with local PvP and Zombie modes. Movement, pursuit and timing are distributed across Web Workers so Phaser's main loop stays focused on rendering and collisions.

JavaScriptPhaser 3ReactWeb WorkersVite

pouv2.netlify.app · Abrir

Open game fullscreen View code on GitHub

How it is implemented

Independent workers process both players' controls, zombie pursuit and the timer.
State travels through postMessage while Phaser keeps rendering and collisions on the main thread.
Includes PvP with WASD and arrow keys plus a 60-second Zombie survival mode.

How I work

The process I used at Handbook, at Agora and with every client.

What I do every week so an operational process ends up running on its own.

Understand the process with its operators

Before writing code I talk to the customer: what gets done, who does it, where it gets stuck and what evidence is needed. That produces prioritized requirements.

Model the full workflow

Stages, rules, handoffs, SLAs and human intervention points. The LLM goes where it adds value, with structured outputs and traceability.

Ship with tests and metrics

CI on GitHub Actions, tests with Playwright, Pytest and Jest, and impact metrics: percentage automated, production errors, load times.

Outcomes

Measurable results per project.

Every metric comes from a specific project, with its context.

Nanti · 2024

40% of support automated

OpenAI-powered flows connected to WhatsApp and tawk.io, running in production.

Agora · 2024-2026

25% fewer production errors

TypeScript, Redux and type-safe architecture across the product layer.

Agora · 2024-2026

20% faster initial load

Scalable component library with Next.js and Tailwind CSS.

Zona Sana · 2023

12% less form drop-off

Complex pharmacy forms redesigned in Next.js, measured across the full flow.

Stack

What I work with.

Six areas: what I use in production and what I have operated before.

Area

AI & automation

OpenAI APIOpenAI AgentsClaudeFunction callingStructured outputsHuman-in-the-loopTensorFlow / Keras

Area

Backend

PythonFastAPIPydanticSQLAlchemyAlembicNode.jsNest.jsExpress.jsHexagonal architecture

Area

Frontend

Next.jsReactTypeScriptTailwind CSSReact Flowshadcn/uiReact QueryReact Hook FormZodReduxAngular

Area

Data & databases

PostgreSQLMySQLRedisApache AirflowETLpandas

Area

Testing & quality

PlaywrightPytestJestReact Testing LibraryMSWPyrightRuff

Area

DevOps, real-time & integrations

DockerGitHub ActionsAWSVercelWebSocketsServer-Sent EventsGmail / Outlook / ZohoWhatsApp

Credentials

Education, certifications and languages.

The formal base behind the practice: software engineering, cloud and UX.

Education

Formal foundation in software and applied programming to design and maintain production systems.

Software Engineering — Universidad Politecnica de Chiapas
Programming Technician — CETIS 138 (2020)

Certifications

Cloud, UX and networking fundamentals that complement full stack practice.

AWS Academy: Cloud Architecting, Cloud Foundations, Cloud Security Foundations and Introduction to Cloud
Google UX Design Specialization (2024)
Cisco: Cybersecurity, Networking and Operating Systems (fundamentals series)

Languages & availability

I worked daily in English, written and spoken, with a US team.

Native Spanish
Professional English: daily use on a US team through 2026
Available immediately for remote and international roles

Verified badges

The most relevant ones for cloud and systems roles; the rest live on my Credly profile.

Contact

Looking for my next team.

Available immediately. I can walk you through the cases and how I built them. Remote AI Full Stack, AI Systems or Product Engineer roles, in English or Spanish.

Send email View LinkedIn

EmailEliseo10208@gmail.com GitHubgithub.com/Eliseo10208 LinkedInlinkedin.com/in/rodrigo-eliseo-garcia LocationMexico City, Mexico