Alonso Silva 

@[email protected]
134 Followers
197 Following
177 Posts
Researcher on Verifiable AI at Nokia Bell Labs | Safran, Berkeley, Inria alumni | Franco-chilean in Paris | Interested about Machine Learning and Data Science
GitHubhttps://github.com/alonsosilvaallende
Alonso Silva Jan 23
If you are (or know of) a Master's or PhD student looking for an internship, I am proposing the subject: 'Efficient Structured Generation with Grammar-Aware Sampling Techniques.' https://www.dropbox.com/scl/fi/7iwfgc4waiszm15urhp1q/Internship_Proposal_2026.pdf?rlkey=3dnmf7nsr5kh9bfx5wwuncwwx&st=taxikz6u&dl=0
If you're passionate about structured generation, feel free to reach out!
Internship_Proposal_2026.pdf

Shared with Dropbox

Dropbox
Alonso Silva Jan 23

itelines got added to the Awesome LLM constrained decoding repo 😊
It’s great to share this space with more established libraries like Outlines, XGrammar, or Guidance.

Link to the Awesome LLM constrained decoding repo:
https://github.com/Saibo-creator/Awesome-LLM-Constrained-Decoding

Link to litelines:
https://alonsosilvaallende.github.io/litelines/

Alonso Silva Jan 19
You can display SQLite database diagrams in @marimo_io using `fastlite` and `graphviz`
Here is a basic notebook to play online:
https://molab.marimo.io/notebooks/nb_ktbJEaXyet6QEUYiXgVf34
Here is my merged PR 🙂
https://github.com/marimo-team/marimo/pull/7787
Alonso Silva Jan 19

.@marimo_io now supports graphviz

Here is a basic notebook to play online: https://molab.marimo.io/notebooks/nb_3xrDJBQEtxscgkKqAo4TCK

Here is the merged PR:
https://github.com/marimo-team/marimo/pull/7787

Alonso Silva Jan 9
Batch processing using transformers and litelines libraries. In this video, I process 900 prompts in 30 seconds with an RTX A4000 with 16GB of VRAM.
https://www.youtube.com/watch?v=7hVUPXxuetk
Batch processing using transformers and litelines libraries

YouTube
Alonso Silva Jan 7
The new litelines release should work much better in marimo notebooks
You can try it in a marimo molab:
https://molab.marimo.io/notebooks/nb_232GhR7P6XJVxz8opD1MWS/app
Here is litelines documentation:
https://alonsosilvaallende.github.io/litelines/
Here is the release changelog:
https://github.com/alonsosilvaallende/litelines/releases/tag/v0.1.4
Alonso Silva Jan 6
New blog post: Force a Qwen model not to use Chinese
https://alonsosilvaallende.github.io/blog/posts/2026-01-06-Force_a_Qwen_model_not_to_use_Chinese/notebook.html
Alonso Silva Jan 3
How is it possible that a 1.7 billion parameter model succeeds where a model with hundreds or thousands of billions of parameters fails?
https://www.youtube.com/watch?v=wmgwuTMauvU
Qwen3-1.7B beating GPT-4o at a lipogram task

YouTube
Alonso Silva Dec 30

The latest release of Litelines supports batch processing for Transformers library.
`pip install --upgrade litelines`

Here is a colab to get started:
https://huggingface.co/datasets/alonsosilva/litelines-notebooks/blob/main/Litelines_Batch_Processing_Multiple_Choice.ipynb

And here is the library documentation:
https://alonsosilvaallende.github.io/litelines/

Alonso Silva Dec 21
My talk, "Processors for Language Models," at PyData Paris 2025 is now available. I discuss my personal project, Litelines, as well as common libraries used to transform unstructured data into structured data, such as Instructor, DSPy, BAML, Outlines, XGrammar, and Guidance.
Here is the link to the video:
https://www.youtube.com/watch?v=VP4IqFpdPec&t=545s
Lightning talks - session 1

YouTube