How to scope a 30-day AI pilot without burning budget
A measurable 2–4 week pilot beats endless demos. Define KPI, available data, out-of-scope and go/no-go before writing code.
Read article →Agents · Automation · Vision · LLM/RAG
Includes biometric systems live at 50+ service points.
I design and ship AI systems into real products and operations, end-to-end. Direct collaboration, modern stack, and measurable production results.
Years of experience
Published cases
Specialized sectors
Named companies / clients
Core offer
AI as the primary offer. Web, mobile and APIs as the delivery capability to ship systems to production.
Agents, automation, RAG and vision/biometrics wired into your products and operations — production-ready.
Ideal for:
Fraud/KYC, back-office, LLM product features
Native Android apps (Java/Kotlin) and hybrid apps (Ionic 7/Capacitor) that work seamlessly.
Ideal for:
E-commerce, delivery services, IoT, enterprise systems
Modern, scalable, and responsive web platforms that drive your business forward.
Ideal for:
Dashboards, CRMs, corporate portals, e-commerce
Robust servers and RESTful/GraphQL APIs to connect your systems.
Ideal for:
System integration, microservices, complex applications
Want a measurable AI pilot?
Book AI diagnosisCompleted projects
Success stories that demonstrate my experience across different sectors and technologies. Each project comes with measurable results and satisfied clients.
Client: Codesa S.A.
The Challenge
Implementar un sistema seguro y rápido de identificación biométrica para validar la identidad de clientes en puntos de atención, reduciendo el fraude y mejorando la experiencia del usuario.
The Solution
Desarrollé una solución integral que combina reconocimiento facial con IA, integrada con Angular en el frontend y Spring Boot en el backend, permitiendo validación en tiempo real con alta precisión.
Technologies
Results
Need something similar?
Book AI diagnosisCompetitive advantages
Direct work, no intermediaries. Proven experience in complex projects and commitment to quality in every delivery.
Direct communication with you, no platform commissions. Greater control and smooth communication.
Career since 2017 at recognized companies (Codesa, El País, Virtual System). Specialization in critical sectors such as transactional and biometric.
Not just a PoC: I wire LLM/vision with Angular/Next, Android and Spring/Node into the real process.
Agile methodology with weekly sprints. Constant updates on your project's progress.
I don't leave you alone after the project. Available for adjustments, questions, and maintenance.
Per project (fixed scope), hourly (competitive rate), or monthly retainer.
Codesa S.A., El País, Virtual System
SENA, Platzi AI, Google Cloud ML
First free consultation within 24 hours
AI methodology
Discovery → Pilot → Production → Improve. Measurable results in weeks, not research papers.
Data, APIs, KPIs and constraints (privacy, legacy, budget).
Free diagnosis · WhatsApp or video call
Scoped delivery with a success metric. Real users or sites.
Usable deliverable · Fast iteration
Secure integration, monitoring, NDA and deploy on your stack.
Maintainable code · Docs
Quality evaluation, API costs and next features.
Optional retainer · Per milestone
Stack for production AI
LLM APIs, Python, evaluation, Angular/Next, Android, Spring/Node and cloud — from idea to deploy.
Proof
Proof via Codesa biometric case metrics and real companies — no invented quotes.
Flagship case
Client: Codesa S.A. · Facial recognition at service points
Facial recognition to validate identity at service points: less fraud, validation in seconds, live across the ops network.
Reduction in fraud attempts
Validation time (was ~5 min)
Service points in operation
I have worked with:
Facing a similar identity, fraud or ops problem?
Book AI diagnosisFirst free consultation
Clean code & best practices
NDA available if needed
Cases with documented metrics
SEO authority
Guides and cases to decide pilots, agents, RAG and identity — linked to solutions and resources.
A measurable 2–4 week pilot beats endless demos. Define KPI, available data, out-of-scope and go/no-go before writing code.
Read article →A chatbot answers; an agent acts with tools and APIs. The right choice depends on the process, not the model hype.
Read article →RAG connects an LLM to your internal knowledge. It shines with docs and policies; it fails on chaotic data or when you need actions, not answers.
Read article →Facial recognition in production: fraud down, validation in seconds — stack, challenges and results.
Read article →Free resource
Use it in a 30–45 min internal meeting. If answers are missing, the next step is a diagnosis — not a big build.
Lead magnet
A 12-question checklist to define scope, data and risks before you invest. Delivered by email.
Start project
Tell me the process or product where you want AI. Reply within 24 hours — or WhatsApp for a quick diagnosis.