AI & Data Science Leader — Enterprise ML, GenAI & Agentic Platforms · Team and Portfolio Ownership
PhD Computer Science · MSc Artificial Intelligence. I lead a 16-person team and a ~€1.3M platform budget at BASF Digital Solutions, own two production AI/ML platforms, and make the portfolio calls that follow — including recommending the closure of one of my own.
Data scientist turned AI/ML leader (PhD Computer Science, MSc Artificial Intelligence) with five-plus years building and running multidisciplinary data science teams and a ~€1.3M platform budget inside a global corporation. I scaled a team from 8 to 16 data scientists, ML/software engineers and solution designers, and owned two production AI/ML platforms covering the full analytics journey from descriptive to predictive, delivering recurring savings above €3.5M.
I own the portfolio calls that follow — what to scale, what to hand over and what to shut down — alongside vendor strategy, demand intake and prioritisation across departments and European markets. I combine hands-on technical depth in ML, LLMs and agentic workflows, still shipping production code solo as co-founder of a live B2B SaaS platform, with life-sciences training in biochemistry, bioinformatics and biomedical machine learning on patient-derived data, and a track record of translating analytics into quantified business cases defended in front of senior leadership.
Four initiatives that show how I run an AI portfolio: what gets built, what gets scaled, what gets handed over — and what gets shut down.
Owned BASF's enterprise computer vision / MLOps platform, with a peer-reviewed library of 70 reusable templates and platform health tracked through active-user adoption and time saved per user. Once mature, I transferred it to its best-fit owner rather than holding on to it.
~€2M recurring savings · 70 templates · ~90% effort reduction for ~100 data scientistsLed a self-service time-series platform giving business users forecasting and decision support without a data scientist in the loop. After an ROI and adoption review, I recommended closing it — a platform I owned — and defended that call with numbers.
~€1.5M savings from the closure decisionTurned an image-labelling bottleneck into a quantified business case presented directly to VP stakeholders, and secured funding for GenAI-powered labelling automation. Ran a proof-of-concept of SIGMA AI and benchmarked competing tools to inform the build-vs-buy decision.
€560K secured · GenAI labelling plus five further innovation initiativesOwned BASF's Ontology Management Framework: the semantic foundation that makes heterogeneous enterprise data interoperable and machine-readable for AI and agentic systems, and the governance model that keeps it alive after launch.
20 bio-ontologies adopted as department standard · GOMO framework publishedOwn a ~€1.3M annual platform budget and a 16-person team (scaled from 8: data scientists, ML/software engineers, solution designers) across BASF's computer vision, MLOps and forecasting platforms, serving the company's global data science community. Accountable for portfolio strategy, vendor decisions and demand prioritisation across three departments and multiple European markets, reporting monthly to Finance and presenting investment cases at VP level.
Owned BASF's Ontology Management Framework (OMF) — enterprise strategy, governance and technology stack (Protégé, RDFlib, Virtuoso, automated curation pipeline) for bio-ontologies underpinning downstream AI and agentic systems, including ontologies authored in-house. Led 5 engineers and 2 external partners, evaluated Stardog as an alternative before retaining and extending the incumbent stack, and communicated ontology strategy and adoption from general-audience sessions up to VP-level stakeholders across the business.
Three principles tested against a real portfolio, a real team and real investment decisions.
Doubled the team through deliberate hiring, individual career paths and coaching of the managers who report into the role — including the hard conversations around underperformance rather than routing around them, because a team's ceiling is set by its weakest link.
Recommended closing a platform I owned once the ROI and adoption data no longer supported it, and transferred another to a better-fit owner — both decisions defended with numbers in front of senior leadership.
Comfortable challenging model choices with ML engineers in the morning and framing the same problem as a business case for executives in the afternoon; this is where most analytics value is won or lost.
vigiaprl.es ↗ Closed beta Access requires invitation credentials.
Co-founded and single-handedly designed and built, end to end, a B2B marketplace platform for a regulated Spanish services sector: Java/Spring Boot backend, React/TypeScript frontend, Keycloak-based identity and access management, geospatial PostgreSQL/PostGIS, document-processing pipeline with OCR, GDPR-compliant data handling, and Verifactu-compliant e-invoicing — from architecture through production deployment.
The scientific track record behind the leadership: first-author publications in Q1 journals and IEEE, industry speaking, editorial work and teaching. h-index 5 · 132 citations.
Development of a fixed module repertoire for the analysis and interpretation of blood transcriptome data
GSAn: a Web server as an alternative to enrichment analysis for annotating gene sets
Presents GSAn, a web server offering an alternative to classical over-representation analysis. Clusters semantic relationships between annotation terms to provide richer, more interpretable gene-set annotations. Still live and in use.
EPPO Ontology: A Semantic-Driven Approach for Plant and Pest Codes Representation
Development of an internal BASF ontology representing EPPO plant and pest codes, enriched with NCBI Taxon data, plus its adoption within BASF's Agricultural Solutions division and the lessons learned.
Ontology Management in an Industrial Environment: The BASF Governance Operational Model for Ontologies (GOMO)
BASF's Governance Operational Model for Ontologies (GOMO), a framework covering the full ontology lifecycle inside a large industrial organisation. Collaboration with the Ontology Engineering Group at UPM.
Federating and querying heterogeneous and distributed Web APIs and triple stores
A federated data integration architecture for an industrial setup, using ontology-based data access to homogenise fragmented Web APIs and triple stores at BASF. Most queries answered in under a second.
A new method for evaluating the impacts of semantic similarity measures on the annotation of gene sets
A novel evaluation methodology for semantic similarity measures applied to gene set annotation, quantifying how measure choice affects downstream biological interpretation.
Deciphering gene sets annotations with ontology-based visualization
An ontology-driven visualization approach for interpreting gene set annotations. Invited talk — travel grant awarded by the Société Française de BioInformatique (SFBI).
Guest Editor, Plants (MDPI, Q1) special issue on bridging the annotation gap in non-model plant species. Kempestiftelserna Fellowship and 2 competitive research grants secured as main author (KSLA, SPPS). Peer reviewer for Microorganisms (MDPI), useR! 2020 and the ISCB Student Council Symposia.
100+ hours of university teaching at master's and bachelor's level (Umeå University, Université de Bordeaux), covering applied functional genomics, databases and computer science fundamentals. Mentored 25+ students across four academic years, plus co-supervision of a master's thesis.
Universidad Internacional de La Rioja (UNIR) · 2023–2024
Université de Bordeaux / LaBRI · Ministerial funding (MESR) · 2016–2019
University of Murcia · 2013–2015
University of Murcia · 2009–2013