Iryna Kondrashchenko

LLM-based coding agents can navigate codebases, edit files, run tests, and iterate with little supervision. Teams have started applying them to ML projects, where the volume of repetitive experimental work makes automation appealing.

In-Browser ML is no longer a toy. You can run Python in WebAssembly, ship interactive demos as a URL, and do on-device ML and LLM inference with no installs, servers, or data leaks. This talk maps what’s possible today, and the real constraints you must design around.

Learn how to reliably evaluate LLM-powered systems in production. Understand how to turn expert knowledge into automated tests, use LLMs as calibrated judges, and combine metrics and human feedback to drive continuous system improvement.

What if you could run real data/ML workflows right in your browsers - sandboxed, with no installation or sending your data anywhere? Such an approach would have tons of benefits: it is easy to distribute, safer by default, and can scale almost infinitely with virtually no infrastructure costs.

Evaluating large language models (LLMs) in real-world applications goes far beyond standard benchmarks. When LLMs are embedded in complex pipelines, choosing the right models, prompts, and parameters becomes an ongoing challenge.

With the rise of LLMs, prompt engineering has become a highly impactful skill in the AI industry. However, manual prompt tuning is challenging, time-consuming, and not always generalizable across different models.

Learn how automated prompt optimization is replacing manual tuning through techniques like bootstrapped few-shot, MIPRO, TextGrad, and practical tools like DSPy and AdalFlow. Understand their benefits, trade-offs, and answer the question: is manual prompt engineering dead, or has it just evolved?

The instruction following and in-context learning capabilities of LLMs make them suitable for tackling many NLP tasks. In this talk, we will introduce [Scikit-LLM](https://github.com/iryna-kondr/scikit-llm "Scikit-LLM"), a rapidly growing, beginner-friendly library that abstracts the complexity...