2025 was the year of artificial intelligence (AI).
Over the long New Year’s weekend, I became seriously interested in large language models (LLM).
I have been developing approaches to working with local models such as llama, phi, and gemma to solve problems of executing instructions and searching for information.
I continued my collaboration with Finam. I have been promoting the use of AI in my department.

In my free time, I started learning Python, a language I had no prior experience with. By spring, I had created multiple code examples about programming large language models, both local and cloud-based. I also gave a talk on artificial intelligence to my colleagues. As one example, I created a programming agent that worked with JIRA and Bitbucket, independently writing code, saving results, and validating the results. I used LangChain and LangGraph, and a combination of local and cloud-based models.

I experimented with creating data sets. In particular, I created my own virtual deputy who was able to answer some questions about my work.
For the UI, I used a JavaScript chat widget, which was created under my supervision entirely using GPT-4, without my direct intervention in the code. This was my first and last vibe coding experience.
I also continued to work on ASP.NET Core Blazor WebAssembly projects in Finam. And by the end of the year, I had been promoted to Chief Programmer.
This year, I created my own knowledge base using Docusaurus.
www.nemiro.net - my CV on GitHub Pages using Jekyll
Knowledge base - my knowledge base on GitHub Pages using Docusaurus + TypeScript + React.js
Aleksey Nemiro
2026-01-03