AI that survives an audit
We take AI into production for regulated enterprises, with results you can measure and governance an auditor can follow. No long-term lock-in.
Founder-led · Marcelo Katogui
Pilar de la Horadada, Alicante, Spain
EU-based · GDPR compliant
KATONOMETRICA S.L.U. · CIF B21631403
Why AI projects stall before production
Models that never reach production
Dashboards without predictive intelligence
Lack of governance and audit readiness
Siloed data that never reaches decision-makers
Most AI projects stall at proof of concept: the people who frame the problem rarely build the system, and the people who build it rarely stay accountable for it. We do both, and every pipeline ships with its audit trail.
That's the gap we're hired to close
Numbers from production
Regulatory reconciliation ran on manual multi-source checks: over 8 hours per cycle, against live submission deadlines. We industrialized the pipeline end to end across the firm's EU compliance workflows.
An internal tool was being rebuilt from scratch for every engagement. We turned it into a multi-tenant platform with regulated data residency, and the client now sells it as a recurring-revenue product.
Reporting arrived in 30+ member-state formats with no unified view. The infrastructure we built now serves the agency's daily reporting under institutional governance, with every run traceable for audit.
Recurring data-quality failures were forcing regulatory resubmissions. We rebuilt the compliance pipeline around validation gates and observability, and downstream errors dropped by 90%.
Primary industries
Financial services
Regulatory reporting, risk dashboards, compliance pipeline infrastructure
Healthcare and life sciences
Clinical analytics, multi-system integration, operational visibility
Insurance
Claims processing infrastructure, underwriting analytics, compliance systems
Public sector
Compliance reporting, data transparency, continental-scale health infrastructure
Selected additional deployments include logistics and agriculture forecasting systems.
(Active clients in Spain and the EU)
Built to pass procurement, audit, and legal review
- EU-based entity (registered Spanish S.L.U., CIF B21631403)
- GDPR-compliant by design, with data processing agreements available on request
- All deliverables in writing with version control and signed acceptance
- Audit trail for every pipeline run: input, transformation, and output provenance
- Infrastructure code owned by the client upon project completion
- SLA response commitments documented in engagement contracts
- No offshore sub-processing of client data without explicit written consent
- Engagement contracts available in English, Spanish, and other EU languages upon request
Forecasting, optimization, and compliance systems in production
Business impact
Forecast accuracy. Predictive models that cut planning error and free up working capital.
Lower operating cost. Model-driven optimization of processes and resource allocation.
Pricing intelligence. Elasticity modeling and dynamic pricing that expand margin.
Compliance automation. Audit-ready pipelines that reduce regulatory exposure.
What we build for you
Forecasting and predictive models
Time-series forecasting, multivariate models, and causal inference, deployed as reproducible pipelines that cut planning error and quantify risk.
Optimization engines
Constraint optimization, resource allocation, and scenario simulation built for operational environments, so decisions get faster and cheaper without breaking regulatory constraints.
Pricing systems
Elasticity modeling and dynamic pricing running as production services. Pricing decisions backed by a model instead of a hunch.
MLOps and governance
Model versioning, experiment tracking, and pipeline orchestration, with the compliance controls regulated enterprises require.
Deployment framework
Diagnostic Assessment (2–4 weeks)
Data readiness evaluation, model feasibility analysis, and prioritized use case roadmap.
Proof-of-Value System
Working prototype deployed in your environment with validated metrics.
Full AI Deployment
Production-grade system with governance, monitoring, and documentation.
Managed AI Infrastructure (SLA-backed)
Ongoing operation, retraining, drift monitoring, and performance guarantees.
Typical timeline: 4–24 weeks
Custom budget after initial diagnostic assessment. No lock-in contracts.
How we work
Direct accountability
Every project is led by specialists with direct experience in quantitative systems for regulated industries. The person who scopes your system is the person who builds it.
AI that runs in production, not in slides
We combine machine learning, time-series forecasting, simulation, and optimization, engineered for secure, governed deployment. If the model doesn't run in production with an audit trail, we haven't done our job.
Verify first, then scale
Every engagement begins with a Diagnostic Assessment (2–4 weeks) to evaluate fit and produce a concrete Proof-of-Value system before larger commitments. You keep the scoped findings, model validation, and working prototype whatever you decide afterwards. No lock-in.
Marcelo Katogui
Founder & Principal
M.Sc. Econometrics (Universidad Torcuato Di Tella). R development contractor at ECDC; consulting roles at Kynetec and Market Fusion Analytics. Specializes in production-grade digital infrastructure for regulated, data-intensive organizations. Background in applied econometrics, forecasting systems, and data integration architecture.
We work with regulated enterprises that need AI running in production. You get direct access to quantitative systems architects on every engagement.
Based in Alicante, Spain. Founded in 2025 by professionals with extensive experience in quantitative systems and production AI for regulated industries.
Registered Spanish S.L.U. (CIF B21631403) · EU/GDPR compliant
Concrete deliverables on every engagement
Diagnostic report
Data readiness assessment, model feasibility analysis, and prioritized AI use cases, delivered as a structured document you keep.
Production pipeline
A versioned, tested, documented ML pipeline deployed to your environment, with validation gates and monitoring built in.
Governance and audit log
Full lineage for every model run: training data, feature versions, hyperparameters, outputs, and drift metrics. Exportable on demand.
Monitoring dashboard
Live visibility into model performance: prediction accuracy, data quality, drift, and SLA breach alerts, deployed alongside the production system.
Runbook and handover pack
Operational documentation covering model retraining procedures, failure modes, recovery protocols, and maintenance runbooks.
Compliance evidence
Structured output for regulatory review: model cards, data flow diagrams, processing records, validation evidence, and AI governance documentation.
Everything we deliver runs in production, under your governance, with the operating guarantees below in writing.
Documented in every managed infrastructure contract.
Scope, outputs, and acceptance criteria agreed before work begins.
Pipeline health and data quality visible before go-live.
Run logs, lineage records, and compliance evidence come standard.
Let's discuss your project
Book a discovery call
Talk through your infrastructure constraints and regulatory context directly with a senior practitioner.
Book a 30-minute callDiagnostic Assessment · 2–4 weeks · No lock-in