Muse Glimmer is gaining attention in the local AI community and is being compared with Qwen’s 27B-class models. In many cases, it is performing better, especially for local coding and agentic workflows. Meta looks strong in the open-model space, and with a few more iterations, models like this could start competing closely with proprietary systems. …

Run Muse Glimmer for Local Vibe Coding with llama.cpp, DFlash, and Pi Read More »

This post was co-authored with Max Silfverberg (Data Scientist, AI Solutions Lead), Antti Hallavo (Lead AI Software Engineer), and Pontus Huotari (Lead Data Scientist). We work at Alma Media, a Finnish digital services, marketplaces and media company. One of our focus areas is developing AI/ML solutions for real estate listing services, where understanding image content plays an important role.  services handle hundreds of thousands of listings a year. Most of those come with …

Why We Fine-Tuned SigLip (And Why That’s Not Always the Right Call) Read More »

  Agentic AI has officially moved from the research lab into enterprise production. In 2026, the AI narrative has shifted dramatically from conversational chatbots — systems that wait for human prompts to generate text — to autonomous AI agents. These systems can plan, execute, and adapt multi-step tasks across external tools, databases, and APIs without …

5 Real-World Use Cases for AI Agents Transforming Industries Read More »

Contents? Let me provide a real, personal example.In this article, I’ll cover:When to fine-tuneWhen to use RAGThe mathematical intuition behind LoRA/QLoRALoRA (Low Rank Adaptation)The core intuition of LoRAWhat this means for practicionersQLoRA (Quantized Low Rank Adaptation)4-bit NormalFloat (NF4)Double QuantizationPaged OptimizersWhy QLoRA instead of full fine-tuning?The technical implementation detailsGenerating high quality training dataWhy 4 different inputs …

How to Fine-Tune an LLM: An End-to-End Guide Read More »

  Building an agent that works in a notebook takes an afternoon. Getting that same agent to survive real traffic, recover from a crash at 3 am, and not leak someone else’s data while it runs large language model (LLM)-generated code is a different job entirely, and it’s the job most teams underestimate. Only a …

5 Tools for Building and Deploying AI Agents in Production Read More »