Welcome to Cloud Native Trento!

Join us for this event and help us shape the world of cloud native technologies in Trento and Trentino Alto-Adige.

An official user community group of the Cloud Native Computing Foundation (CNCF).

What to expect

Vendor-agnostic talks, no sales pitch, from practitioners to practitioners

• 🗣️ Two or Three speakers

• 👥 Meet old and new friends from the community

• 🍕 Free food and drinks afterwards

• ✔️ Just RSVP and book your ticket (it's free, but we need to know how much food and drinks to buy)!

Agenda

• 18:15 Doors open and welcome

• 18:30 Marco Franzon - ⁠Secure by Default: Docker Sandboxes as a New Layer of Isolation & Docker Compose Meets AI: From Application Stacks to Model-Driven Development

• 19:00 Daniele Contessi - MLflow in Practice: From Experiment Tracking to Model Serving in Production

• 20:00 Food and drinks

Speakers

Marco Franzon is a Docker Captain, Software Engineer at eXact lab (specialized IT company in high-performance computing and software development), and co-founder/CTO of Dualistic, a startup focused on digital twin technology. He is an advocate for containerization, DevOps, and reproducible workflows, focusing on Docker extensively in MLOps, machine learning deployment, and high-performance environments to bridge research prototypes to scalable production systems. Marco actively contributes to the Docker community through speaking, content creation, and open-source work, sharing expertise in containers, Python, AI, and modern development.

Daniele Contessi holds a PhD in Physics awarded jointly by the University of Trento and the Universität zu Köln, with a thesis on the application of traditional and unsupervised Machine Learning techniques to the detection of phase transitions in many-body systems. Data Scientist and AI Architect, he is co-founder and CTO of FluidData, an innovative start-up specialised in the development of artificial intelligence algorithms and data-driven solutions for industry: from smart building energy optimisation to document automation through computer vision. He is also co-founder of TravelBrain, where he leads the development of dynamic pricing systems, demand forecasting and revenue-marketing optimisation for the tourism and hospitality sector, with publications in international journals.​​​​​​​​​​​​​​​​

Sessions

Secure by Default: Docker Sandboxes as a New Layer of Isolation by Marco

AI coding assistants can generate working code in seconds but that speed introduces a new risk: developers are increasingly running code they did not write, do not fully understand, and have not reviewed. Whether it’s a script, a dependency, or an entire service scaffolded by an AI tool, executing AI-generated code directly on the host machine quietly expands the attack surface of the local environment.
This talk introduces Docker Sandboxes as a “secure by default” way to run AI-generated and experimental code without exposing your system.

Docker Compose Meets AI: From Application Stacks to Model-Driven Development by Marco

With the rise of local AI and Docker Model Runner, Compose is evolving into something more powerful: a declarative layer that unifies traditional application services with AI models running side by side.
This talk explores how Compose is becoming the orchestration glue between APIs, databases, frontends, and local AI models. We’ll demonstrate how developers can define complete, model-driven application environments in a single compose.yaml.

MLflow in Practice: From Experiment Tracking to Model Serving in Production by Daniele

Managing the lifecycle of a machine learning model, from initial experimentation to production, is one of the most underestimated challenges in building AI products. In this talk, we'll explore how MLflow, an open-source platform (Apache 2.0), can become the backbone of a pragmatic MLOps workflow, even for small teams.

Drawing from real-world projects, building energy optimization, signature recognition, and tourism booking forecasting, we'll show concretely how MLflow helps us track and compare experiments, register and version models across environments, serve models in production, and integrate seamlessly into a modern open-source stack.

The talk will be hands-on, featuring code snippets, aimed at developers, data scientists, and DevOps engineers who want to bring order to their ML workflow without relying on proprietary platforms.

Our Sponsor: FluidData

FluidData is an innovative start-up founded in 2023, specializing in AI, Data Science, and Machine Learning solutions for business process optimization. We primarily focus on dynamic pricing algorithms for hotels and hospitality, energy optimization for smart buildings, document automation with computer vision, logistics optimization, and the development of customized SaaS solutions. Our team is composed of researchers with PhDs in Physics and Engineering, and we collaborate with institutions such as Politecnico di Milano, EHL Lausanne, and other European universities.

Check them out at https://fluiddata.net/

Thanks also to Polo Tecnologico Trentino Sviluppo for hosting us.

🌐 Stay connected with the community on LinkedIn or join the conversation in the CNCF Slack at #cloud-native-trento.