Soverius AI

Welcome back!

We hope you had a relaxing summer, enjoyed some time away, and returned with fresh energy. Now, it is time to get back to work - and we have plenty to share.

We were busy over the summer as well.

Two upcoming Packt events

We have two upcoming events with Packt, with more planned for the future.

The first is the free webinar Local Models for AI-Assisted Coding on Tuesday, 8 September. The second is Hands-On Agentic Frontends with AG-UI and CopilotKit, a three-hour workshop on 16 September in which we will build agentic frontend features using Angular and React.

We'll also notify you about future workshops.

Conversations with the community

We were also guests on Armen Vardanyan’s livestream, where we had the opportunity to discuss our work and exchange ideas with the wider developer community.

Before Armen's livestream, we joined Ilyoskhuja Ikromkhujaev for an AI Fireside conversation about context engineering and the ngTunisia community for a session on Loop Engineering with Local Models.

Watch our Angular Space conversation →

Watch the AI Fireside conversation →

Two new articles

We published two practical articles on the Soverius AI blog.

Getting Started with Local LLMs for Software Development opens a new series on using local models in everyday software engineering. It explains the hardware you need, how model size, quantization and context affect memory, and how to connect local models to Claude Code or Codex. Future instalments will cover benchmarks, inference configuration and hardware in more depth.

Building a Local-First AI Standup Bot for Slack shows how to create an asynchronous standup assistant with CopilotKit Channels, Mastra, a local Gemma model and SQLite as the source of truth.

Read both articles on our blog →

Soverius AI in MLCon Magazine

We are also happy to see two of our articles published side by side in MLCon Magazine 4/2026.

Rainer Hahnekamp’s “LLMs und ihre Grenzen” explains the foundations of language models, agents, and context.

Murat Sari’s “Wissen im Tab” demonstrates a fully client-side RAG solution built with Angular, Transformers.js, WebGPU, and Gemma 4.

Together, the two articles approach AI Engineering from complementary directions: understanding how the technology works and putting it into practice.

Explore MLCon Magazine 4/2026 →

Meet us at an upcoming event

Here is where you can find us over the coming months:

17 September in Berlin: “Lokale LLMs: Sind sie den Aufwand wert?” - entwickler Summit
26-27 October in Munich: International JavaScript Conference
28 October in Las Vegas: AI Dev Craft
5 November in Munich: Java-Entwicklung mit lokalen AI-Modellen - W-JAX

If you are attending one of these events, come and say hello.

Quantized AI News returns

We are still working on a special summer edition of Quantized AI News.

It will bring together the most important AI developments from the summer, add the latest news from publication week, and - most importantly - explain what these developments mean for people building real AI systems.

It is a larger edition than usual, but there is a lot to catch up on. Watch this space.

Until then, welcome back - and we hope to see you at one of our upcoming webinars or events.

Soverius AI · Austria

Legal notice (Impressum) · Privacy Policy

hello@soverius.ai

View in browser · Unsubscribe · Email preferences