Skip to content
Not what you were looking for?I have a business
jdanielgh

Open to new opportunities

Jesús Daniel Gómez Hernández

Product AI Engineer · Node.js and TypeScript

I build backend systems and full-stack products end to end — from requirements to production support — across logistics, fintech, insurance and media. I apply that same rigour to AI: RAG in production, LLM applications and agentic development, measured with regression evaluation and cost per run.

Based in
Dijon, France
Experience
8+ years building production software
Present
Limestone Digital
Hiring
Spanish digital nomad visa · Self-employed, invoices B2B
Email meDownload CVSee work for businesses
GitHub — jdanielghLinkedIn — Jesús Daniel Gómez Hernández
Retrato de Jesús Daniel Gómez Hernández

How I work

Three things I do well

01

Backend and APIs

Node.js and TypeScript microservices, OpenAPI-first contracts, Kafka messaging and Azure deployment. Systems other teams consume without having to ask how they work inside.

02

Product end to end

From the ambiguous requirement to production support: data model, API, interface and the product decisions in between. I know which of the three options costs least to maintain.

03

Applied AI, with numbers

RAG with pgvector and embeddings, LLM applications and agentic development. Contextual chunking, metadata filtering, a regression evaluation that checks the answer is grounded in what was retrieved, and cost and intent measured per run. Not demos that work once.

Projects

What I built recently

Conversation Memory Assistant

Personal AI agent

Private repo

An assistant that searches and answers questions about my WhatsApp conversation history in natural language.

  • A 100-question regression framework that validates the model’s answer against the retrieved evidence while tracking intent, latency and cost. The full run costs $1.38; indexing the whole corpus, $8.20.
  • RAG backend in Node.js and TypeScript over 8,200+ messages: contextual chunking, metadata filtering, pgvector semantic search, Voyage AI embeddings and the Anthropic API.
  • PostgreSQL with Drizzle ORM and Docker Compose, so a run reproduces from one machine to the next.

WaveMark

Founder and sole developer

Closed beta

A web application for smart bookmarks on audio: upload a file, navigate the waveform and save exact moments with timestamp, comment and bookmark type.

  • Designed, built and operated end to end with React, TypeScript, Supabase/PostgreSQL and Cloudflare Workers.
  • Privacy-first architecture: the audio never leaves the user’s device, only metadata is synchronised.

create-claude-base

Open source tool

Public

A reusable Claude Code base — rules, subagents, hooks and MCP presets — for starting new projects without repeating the same decisions.

View on GitHub

Track record

Where it was tested

  1. Apr 2024 — Currently

    Product Engineer · Limestone Digital

    Client: OneRail, US last-mile logistics platform

    • End-to-end product features across Node.js/TypeScript microservices and React applications: deliveries, authentication, carrier integrations and multi-dispatch workflows.
    • OpenAPI-first APIs and shared contracts in Express, Exegesis and TypeScript.
    • Azure end to end: Service Bus, Event Hubs, Cosmos DB, MySQL, Redis and Blob Storage.
    • Production issues investigated through code, PRs, Datadog and observability tooling, down to the tested change shipped via CI/CD.

  2. Jun 2022 — Dec 2023

    Node.js SSR Developer · Globant

    ViX (streaming, US market) and Laureate (education)

    • Node.js/NestJS services and serverless workflows on GCP — Cloud Functions, Pub/Sub, Cloud Tasks — handling metadata for 80,000+ VOD assets.
    • GCP infrastructure, IAM and VPN, with deployment reliability and observability tracked in New Relic.
    • System archaeology across several university implementations to design a unified admissions platform.

  3. Oct 2018 — Jun 2022

    Advanced Backend Developer · IAS

    SURA (insurance) and Bancolombia (banking)

    • From requirements analysis and technical design through to the first MVP of SURA’s enterprise insurance sales platform.
    • Real-time quotations and end-to-end policy purchases: 30% less dependency on manual sales and 40%+ improvement in operational efficiency.
    • The platform scaled to 100,000+ digital policy transactions.
    • Legacy Bancolombia banking modules modernised into Java/Spring Boot services.

  4. Jan 2017 — Oct 2018

    Java Developer · PersonalSoft

    Legacy banking modernisation

    • IBM iSeries (AS/400) modules migrated to Java 8 services, modularising the system without losing business functionality.

Industries

  • Logistics
  • Fintech
  • Insurance
  • Media
  • Education

Education

  • MSc in Systems EngineeringPolitécnico Colombiano Jaime Isaza Cadavid
  • Systems TechnologySENA

Languages

  • Spanish — native
  • English — B2
  • French — basic
The same track record, in more detail, is on LinkedIn
Languages and frameworks
Node.jsTypeScriptExpressNestJSReact
Data
PostgreSQLMongoDBRedispgvector
Infrastructure
DockerKafkaAzureGCPAWS
AI
RAGLLM applicationsClaude Code / agents

Contact

Let’s talk

Open to Product AI Engineer, AI Engineer and Backend & AI roles, remote or hybrid from Dijon. I reply the same day.

jesus.gomezher@gmail.com
Download CVGitHubLinkedIn