Gayan de Silva, PhD
Enterprise AI Executive · Professor · EU AI Act Compliance

Gayan de Silva, PhD

Professor of AI · SRH Hamburg Enterprise AI Executive PhD · Machine Learning · 2012 EU AI Act Compliance
20yr
AI/ML in production
45+
Enterprise clients
17
Countries deployed
€20M
Cost savings
IBMLead IT Specialist · AI/ML
CiscoLead Research Scientist
4FinanceHead of Data Science · 17 Countries
Zeta GlobalSenior Director, Data Science
SRH HamburgProfessor of AI · 2025–Now
GDPR18 months early · 17 countries · zero issues
How I work with organisations

Four ways to engage

EU AI Act Technical Audit

Real engineering tests on production AI systems. Calibration (ECE), distribution shift (PSI), explainability (SHAP), adversarial robustness. Every finding mapped to the exact article.

Board & C-Suite Advisory

Structured briefings that translate EU AI Act obligations into board-level decisions, governance responsibilities and a defensible compliance posture your executives can own.

Executive Education

Half-day workshops for leadership teams. Not AI theory — the specific decisions that determine whether an AI programme is compliant, scalable and governable.

Compliance Strategy

End-to-end EU AI Act readiness planning. From system inventory and risk classification through to remediation roadmaps your teams can execute against.

Career

20 years building production AI/ML

2025 – Now
Professor of AI & Program Director
SRH University Hamburg
Leading the MSc in Applied Data Science programme. Teaching Responsible AI, Applied Machine Learning, NLP, and AI Entrepreneurship. Running the UAGF research programme on AI governance automation.
2019 – 2024
Senior Director, Data Science
Zeta Global · Prague
Led data science across 45+ Fortune 500/1000 accounts. Built and managed AI systems processing 200M daily predictions. Implemented model auditing, human-in-the-loop decision frameworks and distribution shift monitoring across US and European operations — years before EU AI Act made them mandatory.
2016 – 2019
Head of Data Science
4Finance · 17 Countries
Built and led data science across 17 country operations simultaneously. Credit scoring, fraud detection and risk management systems processing millions of financial decisions daily. Implemented GDPR across all 17 markets — technical controls, data processors, legal alignment — 18 months ahead of the enforcement deadline with zero compliance issues.
2013 – 2016
Lead Research Scientist
Cisco Systems
Applied machine learning research at the intersection of network security and AI. Published peer-reviewed research on formal analysis of network security properties — foundational work that directly informs current AI robustness and adversarial testing methodology.
2007 – 2013
Lead IT Specialist · AI/ML
IBM
Enterprise AI and machine learning systems for global clients. Completed PhD in Machine Learning during this period. Laid the technical foundations in production AI deployment that underpin two decades of enterprise practice.
2003 – 2007
Technical Manager
Ceylinco Internet Services · Sri Lanka
Technical leadership at Sri Lanka's first ISP — serving 15,000 individual subscribers and 100+ corporate customers. Designed and built one of the region's earliest colocation and dedicated hosting data centres, predating the era of cloud infrastructure. Built foundational expertise in large-scale network operations, data centre architecture and enterprise IT service delivery.
Industry experience

Deployed across regulated sectors

Financial Services & FinTech
Credit scoring, fraud detection, risk AI across multiple regulatory jurisdictions
Cybersecurity
Threat detection, anomaly classification, behavioural AI at enterprise scale
Marketing Technology
Personalisation engines, customer segmentation, 200M+ daily predictions
Manufacturing
Data collection, EU compliance architecture for European supply chains
Energy & Utilities
Predictive maintenance, grid optimisation and operational AI in regulated energy environments
Logistics & Supply Chain
Demand forecasting, route optimisation, warehouse AI at enterprise scale
SRH University Hamburg

Teaching what I have built

MSc Applied Data Science and AI
Responsible AI & EU AI Act

Ethical frameworks, bias detection, fairness metrics, transparency and accountability — directly aligned with EU AI Act Articles 9, 10, 13 and 14. Students build real compliance frameworks.

MSc Applied Data Science and AI
Applied Machine Learning

Production ML systems, MLOps, model deployment, monitoring and lifecycle management. From research prototype to enterprise production — the full delivery arc.

MSc Applied Data Science and AI
Natural Language Processing

Foundational linguistics to transformer architectures. Production NLP systems and their cross-industry applications, including explainability requirements under EU AI Act Article 13.

MSc Applied Data Science and AI
AI Entrepreneurship

Building AI-first companies, product strategy, go-to-market for AI solutions, and navigating EU regulation as a competitive advantage — not a constraint.

Research Programmes

Research that informs practice

An integrated research programme producing interlocking components of a complete enterprise AI governance stack. Every advisory engagement draws on live research — not desk reviews or theoretical frameworks. When I advise a CTO on AI governance, the tools have been validated on real systems across Finance, Retail, Energy and NLP domains.

AAA · Audit Automation
Agentic AI Auditor — Automated EU AI Act Conformity Assessment

A 6-agent LangGraph pipeline that executes a full EU AI Act conformity audit automatically — ingesting AI system documentation and model artefacts, running a 6-phase protocol, and producing a compliance report in under 2 hours. Validated on 3 real AI systems across Finance, Retail and a live production environment.

UAGF-XAI · Explainability
Four-Layer Explainability, Fairness, Uncertainty & Drift Toolkit

A unified Python toolkit integrating SHAP explainability, fairness metrics, conformal prediction uncertainty intervals and drift detection — automatically selecting the minimum sufficient evidence set for each EU AI Act risk tier. Directly satisfies Articles 13 and Annex III requirements.

ADIF · Executive Intelligence
AI Decision Failure Archetypes & Executive Competency Model

NLP and ML analysis of 200+ FTSE 350 and DAX 40 earnings call transcripts to identify seven recurring senior leadership AI decision failure archetypes. Delivers a board-ready diagnostic quantifying CXO AI decision risk and producing a competency gap roadmap.

ALCSM · Talent Intelligence
AI Leadership Credential Scoring & Gap Analysis

NLP pipeline analysing 500 LinkedIn profiles of AI title-holders against a six-dimension credential scoring model, cross-referenced with 200 job postings. Enables organisations to audit their AI leadership bench before committing to governance programmes that depend on leadership capability.

Academic credentials

Education & patents

Education
PhD · Network Security & AI/ML
Brno University of Technology · Czech Republic · 2012
MSc · Information Technology
Keele University · United Kingdom
BSc (Hons) · Electronics & Telecommunication Engineering
University of Moratuwa · Sri Lanka · Top-ranked A-Level, National Mathematics Olympiad finalist
Patents
Consumer Sentiment Analysis for Selection of Creative Elements
US12073438B2 · Granted
Automated Optimal Threshold Selection for Record Linking Using Probabilistic Matching
Application 17/735,093 · Filed 2022
Automated Data Source Weight Selection for Customer Segmentation Clustering
Application submitted 2024
Work together

The next step is 15 minutes

No pitch. Five questions about your AI systems and compliance posture. You leave with clarity on your exposure regardless of whether you engage me.