I flew to Paris to attend PyTorch Conference Europe as a listener. Two days focused on talks around open source and what’s happening across the industry. What surprised me: remarkably little hyp...

I spoke at OpenTalks.AI in Belgrade, Serbia. My talk covered the end-to-end journey of building production-ready adapters for on-device language models, from tuning through evaluation to real-world...

On June 18, I gave an online talk at the Budapest Data+ML Forum. The event is a solid, mid-sized conference that draws practitioners from Central Europe; it is a good place to exchange technical ex...

In the past few years prompt engineering has emerged as a critical technique for optimizing the performance of large language models (LLMs). For example, CoT with a help of reasoning models (e.g.,...

A technical blog about Phi Silica – Microsoft’s on-device language model for NPU-powered Copilot+PCs. The blog covers PTQ, GQA, KV caches, and strategies to increase context length under the con...

I am happy to present CoAT: Corpus of Artificial Texts, a robust and a general benchmark for artificial (GenAI) text detection. It is published with a help of Cambridge University Press. The wor...

I am fond of the Western attitude towards a gap year. There exist travel programs for those who have obtained a BSc and wish to take a break before pursuing their MSc. In my case, the gap year tu...

The MVP for a model serving cloud-native solution. At my previous role, we actively set up a production environment for deploying hundreds of machine learning models. I want to talk about the cent...