Alexander Amin
Principal AI Engineer @ ML6
Google Developer Expert · Cloud AI
Based in Amsterdam.
About
About Me
I'm Alex, Principal AI Engineer at ML6 and a Google Developer Expert in Cloud AI, based in Amsterdam. For over 12 years I’ve built production-grade software, and these days that mostly means getting AI systems past the demo and into real operations.
At ML6 I lead AI engagements for enterprise clients, from architecture to delivery. Alongside that, I architect Marcus, a native AI legal intelligence platform built for Spanish-speaking civil law jurisdictions in Latin America.
As a GDE I speak regularly at conferences and Google Developer Group events across Europe and Latin America, and I lecture at Frankfurt University of Applied Sciences. Alongside German and English, I speak Spanish, Portuguese, Dutch, and Dari.
Earlier: tech partnerships at ARD Online, acqui-hired to Hackerbay, CEO at High5-IT.
Lecturer
Frankfurt University of Applied Sciences
Speaker
Google Cloud Summit DACH, DevFests, and more
Partnered with
Deutsche Bahn, ARD, Siemens, Vodafone, Google
Work
Featured Projects
Marcus
Architecting Agentic RAG and AI Legal Intelligence for Civil Law Jurisdictions.
Marcus is a native AI legal workspace purpose-built for Spanish-speaking civil law jurisdictions. Moving away from rigid prompt engineering, the platform relies on an autonomous, multi-turn agentic routing loop that evaluates user intent, invokes specialized tools, and runs isolated self-review loops to eliminate hallucinations. The core engine utilizes a custom parent-child retrieval schema in PostgreSQL/pgvector with 3072-dimensional Gemini embeddings to map broad legal concepts to granular, verifiable citation trees. From high-throughput asynchronous bulk extraction grids for corporate due diligence to programmatic task pipelines, Marcus bridges the gap between raw statutory corpora and practical legal operations, fully integrated natively into Microsoft Word via Office.js.
Gemini Kicks & Llama Hoops
A Comparative AI Analysis: Pitting Google Gemini vs. Meta Llama for Real-Time Video Highlight Generation.
To test the real-world capabilities of leading AI models, I built two parallel projects that automatically analyze and clip highlight moments from full-length sports games. This exploration into multi-modal AI showcases a practical application of Gemini and Llama 4 for automated content creation, directly comparing their performance on a complex, real-world task. Both projects are open-source.
Railmap for OpenRailwayMap
An Android App for Railway Enthusiasts.
A native Android app offering a clean, mobile-native interface for the extensive data from OpenRailwayMap. Published and available on the Google Play Store.
Speaking
Talks & Presentations
I believe in building a stronger developer community by sharing knowledge and insights from real-world projects. I enjoy breaking down complex topics related to AI and cloud architecture for technical audiences.
"Demystifying Agents"
Understand AI agents like never before by learning to build one from the ground up using just the google/genai library and Firestore. Discover that agents are more accessible than you think and explore how the Agent Development Kit (ADK) can significantly streamline your workflow.
"Building Generative, Streamed UIs with the Gemini API and Vue.js"
Explored how to use structured output from LLMs to reliably generate deterministic JSON, preventing application failures and enabling real-time UI rendering.
"Project Showcase: Llama Hoops"
Demonstrated a practical, multi-modal application of Llama 4 for automated video content creation in a high-pressure hackathon environment.
Interested in having me speak at your event? I'm always open to new opportunities.
Let's TalkContact
Let's Build Something Together
I'm always curious about projects, new challenges and speaking opportunities. If you have an interesting topic, let's connect.