Welcome to My Blog
26 | 27 | 28 |This is my first blog post. I'm excited to start sharing my thoughts and experiences here.
30 | 31 |What You'll Find Here
32 | 33 |I plan to write about:
34 | 35 |-
36 |
- Technology and programming 37 |
- Personal projects and experiments 38 |
- Things I'm learning 39 |
- Random thoughts and ideas 40 |
Why I Started This Blog
43 | 44 |I wanted a simple place to document my journey and share what I learn along the way. This blog is intentionally minimal - just HTML and CSS, no unnecessary complexity.
45 | 46 |Get In Touch
47 | 48 |Feel free to reach out if you want to discuss anything I write about!
49 | 50 |Thanks for visiting!
51 |
11 |
12 |
13 | In Hugging Face we have a culture of filling-in to where needed. At the beginning of the year there was a hype around using vision language models for agency. I have integrated vision capabilities to smolagents, we planned everything on how memory should work, how use cases like agentic browsers need to be written etc, it was super fun. Read more about it [here](https://huggingface.co/blog/smolagents-can-see). Similarly, I worked with TRL team to have more alignment methods for vision LMs ([read more](https://huggingface.co/blog/trl-vlm-alignment)).
14 |
15 | We have written two major blogs around vision LMs this year, you might want to check them out. Last year we were trying to enable vision LM training in TRL, and to launch this, we have written a [blog](https://huggingface.co/blog/vlms) on vision LMs and how they work. This year we have written a follow up [blog](https://huggingface.co/blog/vlms-2025) on the paradigm changes (spoiler: it's a lot!). This year it was raining OCR models, and we were constantly receiving the question "which one is the best?" so we have written [a long post](https://huggingface.co/blog/ocr-open-models) on OCR landscape to help people get started easier.
16 |
17 | I have written a book about vision language models with my friends in SmolVLM 📖
18 |
19 |
21 |
22 |