Curriculum vitae

I make AI work in production · Evals, observability, agentic development and cost-efficient scale · AWS Certified Professional · 17+ years.

Download PDF

Madrid, Spain · Email · LinkedIn · GitHub

SUMMARY

I make AI work in production. I am an AWS-certified engineer with 17+ years in software engineering. At Aircall I build the AI behind call summarization: real-time transcription and LLM features on serverless microservices. The platform handles over a million calls a day at +99.9% reliability. My job is to make that AI predictable, in quality and in cost: evals as CI gates, traces on every call. I am also the founder of Deepgent, a multi-tenant AI chatbot SaaS that is live in production. Before that, I was a CTO and a startup co-founder. I build my products end to end with agents, under spec-driven development and a zero-tolerance CI.

EXPERIENCE

Jan 2026 – Present

Founder

Deepgent — deepgent.net (Remote)

Built a multi-tenant AI chatbot SaaS solo (TypeScript, Python, AWS Bedrock): agentic tool calling with fail-closed auth, identity verification (KBA + OTP), circuit breakers, and per-tenant tool whitelisting.
Built a RAG knowledge base with layered prompt architecture and company-scoped isolation. Added an LLM eval framework (grounding, honesty, LLM-as-judge) as CI quality gates, and an MCP server for AI-assisted admin ops.

Mar 2025 – Present

Senior Software Engineer & AI Architect

Aircall — Madrid (Hybrid)

Designed an AI-powered Call Summarization platform that processes 10K+ daily transcriptions with +99.9% reliability across serverless microservices, for thousands of B2B customers.
Shipped Company Context Knowledge Graph and Prompt Optimizer: streaming LLM features that let customers enrich and customize AI-generated insights in real time.
Introduced LLM evaluation practices with Braintrust benchmarks as quality gates for model migrations, and full observability with Datadog.
Delivered Async Transcription end to end: a new microservice from scratch to production in ~3 weeks (Lambda, SQS, Step Functions, Aurora PostgreSQL), with a public API and webhooks.
Pushed AI adoption across the company: AI Guild standards, an internal SDLC agent, CDO automation, and a company-wide AI Showcase.

Apr 2022 – Mar 2025
3 yrs

Lead Cloud Software Engineer

hireful — Remote

Led migration of legacy apps to cloud-based, event-driven architectures using hexagonal architecture, Node.js, NestJS, Lambda, DynamoDB, Fargate and CDK.
Built CI/CD pipelines with AWS CodePipeline, GitHub Actions, and Docker, with automated test gates for B2B SaaS delivery.
Designed cloud-native microservices with API Gateway, Fargate, and MongoDB, and kept technical decisions in line with product strategy.

Mar 2021 – May 2023
2 yrs

Strategist Product Engineer

Sporttrait — Remote (Part-time)

Defined technology strategy and designed full B2C SaaS infrastructure with Next.js, AWS EventBridge, and containerized deployments on Kubernetes and Docker.

Sep 2018 – Mar 2022
3.5 yrs

Software Architect / Technical Lead

DCSL GuideSmiths Spain — Madrid

Led digital transformation projects for B2B clients. Managed squads of 2–4 people that delivered React.js, AWS, and React Native solutions.

Feb 2018 – Aug 2018

CTO & Technical Co-Founder

Athletika — Alicante

Built the MVP from scratch during the seed phase (React.js, React Native, GCP/Firebase). Played a central role in the business and investment processes.

Jun 2017 – Feb 2018
9 mos

Mobile Lead Developer

everis — Alicante

Led hybrid mobile development for a banking group: customizable apps with security compliance and test automation. Acted as technical reference and internal trainer.

Oct 2014 – May 2015
8 mos

CTO

MedBravo — Alicante

Built a HealthTech SaaS platform for cancer-patient referral to clinical trials. MedBravo was a FICHe finalist at Health 2.0 Europe.

Mar 2009 – Oct 2014
5 yrs 8 mos

Software Engineer

Indra — Madrid

Developed anti-collision algorithms for UAVs in non-segregated airspace (C, C++, Monte Carlo) on the MIDCAS project for the European Defence Agency.

FEATURED PROJECTS & WRITING

Kaiord: an open-source, local-first training platform. AI agents wrote every line, behind a zero-tolerance CI: a TypeScript SDK, a CLI, and an MCP server in the official registry.
Technical writing on evals, agentic development, and observability for AI pipelines.