
LiteLLM is the known option. agentgateway is the open one.
Comparing LiteLLM and agentgateway for enterprise AI platforms: licensing boundaries, security advisories, ops footprints, and open governance.
- Agent
- LLM
- Guide
Transforming AI/ML initiatives into secure, scalable, and high-impact solutions.
I am a freelance AI Architect with a PhD in AI and Computer Security. I design and industrialize AI agents, MLOps platforms, and secure cloud foundations, turning fast-moving research into dependable production capabilities with clear controls, observability, and measurable outcomes.
My work spans strategy and delivery: enterprise agent platforms at Decathlon, European fraud detection for the European Commission, and Android malware research with Google. I have also delivered AI, data, and security initiatives for BNP Paribas, ArcelorMittal, SFEIR, Clearstream, and the University of Luxembourg.
Beyond client work, I serve as an AAIF Ambassador and Luxembourg organizer and contribute to the 33N Ventures Expert Advisory Board. As a Google Cloud Professional Cloud Architect, I bring a pragmatic, security-first approach to systems that must operate reliably at scale.
Building autonomous systems, reasoning engines, and reliable agentic workflows.
Robust deployment patterns on GCP, AWS, Azure, and Databricks for reliability.
Leveraging a PhD background to build secure and trustworthy AI systems.
Translating complex AI capabilities into clear, scalable architectural roadmaps.
Engineering machine learning models and data solutions for enterprise scale.
Building scalable applications, libraries, and tools with modern standards.
A track record of engineering scale, reliability, and security for industry-leading institutions.

Decathlon
Design and implement enterprise-scale Agents and MLOps platforms for AI/ML industrialization.

European Commission
Contributed to Arachne, the European fraud detection system to ensure financial integrity.

ArcelorMittal
Trained and optimized AI/ML models for steel price recommendations in industrial markets.

BNP Paribas
Supervised the development of advanced transformer models for banking applications.

Collaborated with the Android Security team to enhance malware detection and characterization.

SFEIR
Delivered data engineering, technical consulting, recruitment, and pre-sales.

University of Luxembourg
Conducted AI security research and taught AI/ML, Big Data, and Android development.

Clearstream
Selected and configured a scalable SIEM solution to detect and respond to security incidents.
The Linux Foundation
Selected to serve as an AAIF Ambassador and help grow the open agentic AI community.
33N Ventures
Contributing AI Agents, MLOps, and security expertise to 33N's venture advisory network.
AAIF Community Luxembourg
Organizing Luxembourg's local practitioner community and events around agentic AI.
Validated capabilities from leading foundations, cloud providers, and academic networks.
The Linux Foundation • 2026
Google Cloud • f86b41ffd74d4b50947e69876a9274af
Google Cloud • 84c1e188f1054fb8bd53af78060df688
Databricks • 61461287
Microsoft Azure • 992564946
Neo4j • cd64002c-63a3-4968-87dd-d6cca204a5cd
Google • WLU4DBPSQ4B5
Google • EPZ3WQFC423E
Google • YSNPABSMV6JL
Google • V492QQ4JJKEB
Udemy • UC-XALJEH7G
Udemy • UC-5FM0CC9S
DeepLearning • LQ4GHWJ6URBS
Google (Udacity) • PV7A7EAA
Stanford (Coursera)
Open-source repositories and educational deep dives created to upskill engineers worldwide.
Kickstart your MLOps initiative with a flexible, robust, and productive Python package.
Start building and deploying Python packages and Docker images for MLOps.
Learn to create, develop, and maintain a state-of-the-art MLOps code base.
Technical deep dives into the Black Box of AI Agents and emerging autonomous systems.
Brief, high-signal deep dives into the core architecture and inner workings of AI Agents.
Bridge the gap between robust software engineering and cutting-edge data science.
Bring deep AI architecture, MLOps, and security expertise to your next high-impact initiative.
Engage me to assess, design, and scale AI initiatives — from agent platforms and MLOps to security and operating models.
Book a paid 1-hour session to discuss your projects: upskilling, career mentoring, architecture review, brainstorming, and more.
Field notes on AI agents, MLOps, cloud architecture, and security — written from delivery, not theory.

Comparing LiteLLM and agentgateway for enterprise AI platforms: licensing boundaries, security advisories, ops footprints, and open governance.

AI coding agents amplify repository friction. Here is how we redesigned an MLOps codebase around deterministic single-command gates, SQLite isolation, and machine-readable AGENTS.md guardrails.

Running a remote Model Context Protocol (MCP) server behind multiple instances presents immediate infrastructure challenges. Because the…

As AI agents transition from isolated developer experiments into production-scale workflows, engineering organizations are waking up to a…

Most coding-agent tasks have one right answer hiding behind a workflow. “Add password sign-in.” “Fix the off-by-one in the pagination…

A coding agent is only as useful as the spec you give it. That sentence sounds like a truism until you watch it play out for an afternoon.