Senior Data Engineer · GenAI Solutions Architect

Antonio Rocamora Manteca

I modernize data platforms, lead engineering teams and build AI workflows with evaluation and traceability.

Download CV
Antonio Rocamora Manteca

About me

I've spent 17+ years building data platforms across banking, insurance and manufacturing. At NTT DATA, I lead a 70+ person team, designing pipelines on AWS and applying generative AI to the way we work.

When I'm not working, you'll find me running marathons, sketching architecture in ink, and writing about data engineering and AI at El Rincón de Kakashi.

Selected work

Enterprise data Professional experience

BBVA

Data engineering & technical leadership

From critical banking applications at BBVA IT to cloud modernization and generative AI through NTT DATA.

Modernization
Data integration, cloud migration and process automation.
AI & technical standards
AI adoption across engineering, architecture guidance and engineer training.
  • Data integration
  • AWS
  • Generative AI
  • Technical leadership

Leadership at scale

70+

people in the team I lead at NTT DATA, across Spain and Mexico.

  1. –present NTT DATA · Technical authority

    Data Integration, Big Data and AI adoption for the BBVA account.

View earlier roles
  1. – NTT DATA · Cloud modernization

    Cloud migration, hyperautomation and engineer training.

  2. – NTT DATA · Project leadership

    Data integration projects, stakeholder coordination and automation.

  3. – BBVA IT España

    Critical applications, SLA support and business intelligence.

  4. – NTT DATA & CGI

    ETL and BI delivery for banking clients, including BBVA.

Data modernization Proof of concept

Ulysses

From PowerCenter to PySpark

I designed and delivered a migration accelerator that combines metadata extraction with generative AI to turn existing data pipelines into PySpark and Redshift SQL code.

PoC result

Generated code required no manual corrections in the delivered proof of concept.

  • Python
  • AWS Bedrock
  • RAG
  • PySpark
  • Redshift

Migration workflow

  1. Extract metadataPowerCenter XML · Python
  2. Generate with contextAWS Bedrock · Retrieval-augmented generation
  3. Produce validated codePySpark · Redshift SQL
Applied AI Internal application

Inari

Harness engineering for AI evaluation

A desktop application that assesses CVs against a vacancy. I built a controlled pipeline around the models, with evidence checks, scoring rules and an auditable record of each run.

Versioned prompts
Keep model instructions identifiable as the application evolves.
Regression evaluations
Use reference cases to detect changes in scores and evidence checks.
Traceable execution
Inspect model calls, responses and latency to understand each result.
  • Python
  • PySide6
  • Harness engineering
  • LLM evaluation

From CV to human review

  1. Evaluate the fitAssess mandatory and optional skills, one candidate at a time.
  2. Verify the evidenceA second model stage checks claimed skills against the CV.
  3. Calculate the scoreApply corrections and calculate the score with rules in Python.
  4. Support the decisionA manager reviews the results and makes the final decision.
Process & quality Product definition

An AI auditor for CMMI

Turning a maturity model into requirements for an AI-powered auditor. My contribution covers both the functional scope of the assessment and its technical definition.

Project goal

Automate the assessment of process maturity against the CMMI model.

  • CMMI
  • Generative AI
  • Product ownership

Scope & responsibilities

My role
Product owner
Functional requirements
What the auditor should assess
Technical requirements
How the assessment should work

Get in touch

Let's talk data & AI.

For conversations about data modernization, technical leadership or AI in engineering, get in touch.

Email me Download my CV