talk-data.com

Platform Engineering : de la jungle DevOps à l’autoroute des développeurs.

Depuis plus de dix ans, DevOps, CI/CD et Cloud promettent fluidité et autonomie. Pourtant, beaucoup d’équipes subissent encore la complexité des environnements avec sa jungle d’outils et un onboarding difficile. Le Platform Engineering propose une réponse : créer une Internal Developer Platform (IDP), pensée comme un produit, pour offrir aux développeurs une expérience cohérente et en self-service.

En 45 minutes, nous verrons : 🔎 Comment DevOps montre ses limites dans les organisations. 🛠️ Les principes et cas d’usage du Platform Engineering. 🚀 Dans quelles mesures une IDP peut impacter positivement le quotidien. 🧩 Les étapes clés pour bâtir une IDP 💎 Un ensemble de ressources pour alimenter vos projets d’IDP

Vous repartirez avec une vision claire du mouvement et des idées pratiques pour démarrer.

Par Aurélien Maury, CTO de WeScale

Notre partenaire et hébergeur des meetups du 2e semestre 2025 : WeScale

In our upcoming meetup, we'll explore best practices in MLOps, ensuring robust and automated workflows, and discuss the latest advancements in Generative AI for real-world applications. Whether you're optimizing data pipelines, scaling AI models, or navigating the transition from experimentation to production, this event will provide valuable insights from industry experts.

Between presentations, you will have the opportunity of networking and meeting data enthusiasts at the Netlight office, food and drinks will be served.

Agenda:

17:30 - 18:00: Doors open 18:00 - 18:10: Welcome 18:10 - 18:40: Navigating the Intersection of MLOps and GenAI: A Comparative Exploration 18:40 - 19:10: Break 19:10 - 19:40: Building LEGO Castles Instead of Sandcastles: A Tale of Modularity in Data & ML Systems 19:40 - 20:30: Networking

–

Presentations:

Navigating the Intersection of MLOps and GenAI: A Comparative Exploration Per Hedbrant - Consultant, Netlight Martti Yap - Consultant, Netlight

In this presentation, we'll embark on a journey through the evolving landscapes of MLOps and GenAI architectures. Drawing from extensive experience in data engineering and machine learning, coupled with hands-on work in the emerging field of GenAI, we will provide insights into the fundamental differences and similarities between these two domains.

We'll delve into the core components of a mature MLOps platform, highlighting processes like data preparation, model training, and deployment. Then, we'll contrast these with the emerging architecture of GenAI, exploring concepts of observability, guardrails, and model evaluation techniques.

This talk aims to equip you with a deeper understanding of where the focus lies in MLOps—emphasizing operational efficiency and model lifecycle management—and in GenAI—highlighting the demands of AI-driven solutions in production. Whether you're a student, a newly minted professional, or a seasoned expert, this session will provide valuable perspectives on integrating these technologies into your workflow, fostering both operational robustness and creative AI capabilities.

Speakers Bio: Per Hedbrant is a Netlight consultant with a strong background in data engineering and machine learning, currently engaged in advancing Generative AI solutions. Passionate about bridging the gap between traditional ML operations and cutting-edge AI innovations, Per is dedicated to unleashing business value through building AI products and teams.

Martti Yap is a Netlight consultant, with a background in data science and ML. He is currently developing generative AI capacities for industry enterprises. Martti thrives best where evolving business needs meet advanced technological solutions. He enjoys sparking interest and promoting knowledge sharing throughout organizations and teams.

Building LEGO Castles Instead of Sandcastles: A Tale of Modularity in Data & ML Systems Anton Gollbo - Data/ML Engineer, Netlight

Building reliable and scalable machine learning systems is challenging, especially when workflows rely on fragile, tightly coupled scripts and notebooks. These "sandcastle-like" systems—where every component depends on the exact state of the whole—break easily, slowing down iteration and making debugging painful. Without clear modularity, small changes can cause unintended failures, leading to rigid, hard-to-maintain pipelines that don't scale well. To address this, we shift towards a LEGO-like approach, where ML systems are built from small, interchangeable, and testable components. By designing modular pipelines with well-defined boundaries—such as independent data processing, feature engineering, model training, and evaluation steps—our goal is to create flexible and reusable workflows. This talk explores the journey from tightly coupled systems to composable architectures, showing how modular design enables faster iteration, greater reliability, and long-term scalability in ML development.

Speakers Bio: Anton is a consultant at Netlight, bringing extensive experience from data and machine learning projects. His professional journey has taken him through various stages of the data and ML lifecycle, cultivating an interest in constructing systems that are both data-intensive and designed for easy testing and modularity.

–

About the event

Date: March 27th , 17:30 - 20:30 Location: Netlight Consulting AB, Regeringsgatan 25, 111 53 Stockholm. Directions: At the entrance, take the staircase and you will find the reception desk where one of the hosts will welcome you and give more information about the venue. Tickets: Sign up required. Anyone who is not on the list will not get in. The event is free of charge. Capacity: Space is limited to 100 participants. If you are signed up but unable to attend, please change your RSVP by March 26th. Food and drinks: Food and drinks will be provided. Questions: Please contact the meetup organizers.

–

Code of Conduct The NumFOCUS Code of Conduct applies to this event; please familiarize yourself with it before attending. If you have any questions or concerns regarding the Code of Conduct, please contact the organizers.

Brought to you by: • The Enterprise Ready Conference on October 30th — For B2B leaders building enterprise SaaS. • DX — DX is an engineering intelligence platform designed by leading researchers.  • ByteByteGo — Ace your next system design interview. — You may not be familiar with Bending Spoons, but I guarantee you’ve encountered some of their well-known products, like Evernote and Meetup. In today’s episode of The Pragmatic Engineer, we sit down with three key figures from the Italy-based startup: cofounder and CEO Luca Ferrari, CTO Francesco Mancone, and Evernote product lead Federico Simionato. Bending Spoons has been profitable from day one, and there's plenty we can learn from their unique culture, organizational structure, engineering processes, and hiring practices. In today’s conversation, we cover the following topics: • The controversial acquisitions approach of Bending Spoons • How Bending Spoons spent more than $1 billion in buying tech companies • How the Evernote acquisition happened • How Bending Spoons operates and how it organizes product and platform teams • Why engineering processes are different across different products • How ‘radical simplicity’ is baked into everything from engineering processes to pay structure. • And much more! — The Pragmatic Engineer deepdives relevant for this episode: • Good attrition, bad attrition for software engineers: https://newsletter.pragmaticengineer.com/p/attrition  • Healthy oncall practices: https://newsletter.pragmaticengineer.com/p/healthy-oncall-practices • Shipping to production: https://newsletter.pragmaticengineer.com/p/shipping-to-production • QA across the tech industry: https://newsletter.pragmaticengineer.com/p/qa-across-tech — In this episode, we cover: (2:09) Welcome, Luca, Francesco, and Federico from Bending Spoons (03:15) An overview of the well-known apps and products owned by Bending Spoons (06:38) The elephant in the room: how Bending Spoons really acquires companies (09:46) Layoffs: Bending Spoons’ philosophy on this (14:10) Controversial principles (17:16) Revenue, team size, and products (19:35) How Bending Spoons runs AI products and allocates GPUs (23:05) History of the company (27:04) The Evernote acquisition (29:50) Modernizing Evernote’s infrastructure (32:44) “Radical simplicity” and why they try for zero on calls (36:13) More on changes made to the Evernote systems (41:13) How Bending Spoons prioritizes and ships fast  (49:40) What’s new and what’s coming for Bending Spoons (51:08) Organizational structure at the company (54:07) Engineering practices (57:03) Testing approaches (58:53) Platform teams (1:01:52) Bending Spoons tech stack and popular frameworks (1:05:55) Why Bending Spoons hires new grads and less experienced engineers (1:08:09) The structure of careers and titles at Bending Spoons (1:09:50) Traits they look for when hiring  (1:12:50) Why there aren’t many companies doing what Bending Spoons does  — Where to find Luca Ferrari: • X: https://x.com/luke10ferrari • LinkedIn: https://www.linkedin.com/in/luca-ferrari-12418318 Where to find  Francesco Mancone: • LinkedIn: https://www.linkedin.com/in/francesco-mancone Where to find Federico Simionato: • X: https://x.com/fedesimio • LinkedIn: https://www.linkedin.com/in/federicosimionato Where to find Gergely: • Newsletter: https://www.pragmaticengineer.com/ • YouTube: https://www.youtube.com/c/mrgergelyorosz • LinkedIn: https://www.linkedin.com/in/gergelyorosz/ • X: https://x.com/GergelyOrosz — References and Transcripts: See the transcript and other references from the episode at https://newsletter.pragmaticengineer.com/podcast — Production and marketing by https://penname.co/. For inquiries about sponsoring the podcast, email [email protected].

Get full access to The Pragmatic Engineer at newsletter.pragmaticengineer.com/subscribe