Sign in

Alonso Silva

@alonsosilva.bsky.social
253 followers 296 following 92 posts

AI Researcher @ Nokia Bell Labs. Interested on Large Language Models (LLMs), Machine Learning and Data Analysis.

PostsRepliesMedia
Alonso Silva @alonsosilva.bsky.social · 30/04/2026
Ever wanted to render Graphviz graphs in the browser with pure Python? Now you can 🐍🌐 Check out the WebAssembly-powered package gvrender! 👇 #Python #WASM #Graphviz
The image shows the following Python code:
```python
from graphviz import Digraph
from gvrender import render_graphviz

dot = Digraph()
dot.edge("Hello", "World")
dot.edge("Hello", "Name")
render_graphviz(dot)
```
100
Alonso Silva @alonsosilva.bsky.social · 22/04/2026
🐍 New Python package just dropped! Generate Mermaid diagrams directly from Python code. Simple, fast & powerful. 🔥 #Python #Mermaid #OpenSource
Code to generate a Mermaid diagram
Input cell 1: %pip install mermaid-nb
Input cell 2: 
from mermaid_nb import Mermaid

Mermaid(
    diagram="""
flowchart LR
  A[Start] --> B{Decision}
  B -->|Yes| C[Continue]
  B -->|No| D[Stop]
""",
    theme="forest",
    look="handDrawn",
    theme_variables={"nodeTextColor": "#0000ff", "fontFamily": "Caveat, cursive"}
)
120
Reposted by Alonso Silva
Martin Renou @martinrenou.bsky.social · 23/01/2026
Pandas 3.0 was just released yesterday!! 🤘 And guess what? You can already play with it in Notebook.link. I quickly built a Notebook.link link for you to play with it now: notebook.link/@martinRenou...
notebook.link
Notebook Link
044
Alonso Silva @alonsosilva.bsky.social · 23/01/2026
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.' www.dropbox.com/scl/fi/7iwfg... If you're passionate about structured generation, feel free to reach out!
dropbox.com
010
Alonso Silva @alonsosilva.bsky.social · 23/01/2026
Litelines 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 repo: github.com/Saibo-creato... Link to litelines: alonsosilvaallende.github.io/litelines/
Screenshot of Awesome LLM constrained decoding repo.
000
Alonso Silva @alonsosilva.bsky.social · 19/01/2026
You can display SQLite database diagrams in @marimo.io using `fastlite` and `graphviz`. Here is a basic molab notebook to play online: molab.marimo.io/notebooks/nb... Here is my merged PR 🙂 github.com/marimo-team/...
The following code:
```python
from fastlite import database, diagram
from graphviz import Source
from pathlib import Path
from fastcore.net import urlsave

url = 'https://github.com/lerocha/chinook-database/raw/master/ChinookDatabase/DataSources/Chinook_Sqlite.sqlite’
path = Path('chinook.sqlite')
if not path.exists(): urlsave(url, path)
db = database("chinook.sqlite")
diagram(db.t)
```
displays the database diagram.
000
Alonso Silva @alonsosilva.bsky.social · 19/01/2026
marimo @marimo.io now supports graphviz Here is a basic notebook to play online: molab.marimo.io/notebooks/nb... Here is my merged PR 🙂 github.com/marimo-team/...
The following code:
```python
import graphviz

dot = graphviz.Digraph()
dot.edge("hello", "world")
dot.edge("hello", "name")
dot
```
 displays a graph with three nodes: 'hello', 'world',  'name', and two directed edges: from hello to world and from hello to name.
000
Alonso Silva @alonsosilva.bsky.social · 09/01/2026
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. www.youtube.com/watch?v=7hVU...
youtube.com
Batch processing using transformers and litelines libraries
YouTube video by Alonso Silva
000
Alonso Silva @alonsosilva.bsky.social · 07/01/2026
The new litelines release should work much better in marimo notebooks. You can try it in a marimo molab: molab.marimo.io/notebooks/nb... Here is litelines documentation: alonsosilvaallende.github.io/litelines/ Here is the release changelog: github.com/alonsosilvaa...
000
Alonso Silva @alonsosilva.bsky.social · 06/01/2026
New blog post: Force a Qwen model not to use Chinese alonsosilvaallende.github.io/blog/posts/2...
Qwen logo in Chinese.
000
Alonso Silva @alonsosilva.bsky.social · 03/01/2026
How is it possible that a 1.7 billion parameter model succeeds where a model with hundreds or thousands of billions of parameters fails? www.youtube.com/watch?v=wmgw...
youtube.com
Qwen3-1.7B beating GPT-4o at a lipogram task
YouTube video by Alonso Silva
000
Alonso Silva @alonsosilva.bsky.social · 30/12/2025
The latest release of Litelines supports batch processing for Transformers library. `pip install --upgrade litelines` Here is a colab to get started: huggingface.co/datasets/alo... And here is the library documentation: alonsosilvaallende.github.io/litelines/
How to define a logits processor and generate a structured response.How to visualize the selected paths defined by the logits processors.
000
Alonso Silva @alonsosilva.bsky.social · 21/12/2025
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. Link to the video: www.youtube.com/watch?v=VP4I...
youtube.com
Lightning talks - session 1
YouTube video by PyData
000
Alonso Silva @alonsosilva.bsky.social · 11/12/2025
I gave a talk about structured code generation for domain-specific languages: Abstract Syntax Trees (ASTs), Concrete Syntax Trees (CSTs), Deterministic Finite Automata (DFA), Regular Expressions (Regex), Pushdown Automata (PA), Context-Free Grammars (CFGs), outlines, guidance, Georges Perec, etc.
100
Alonso Silva @alonsosilva.bsky.social · 30/09/2025
Here are the starting notebooks I presented at @pydataparis.bsky.social tinyurl.com/litelines-hf And here is the documentation of litelines: tinyurl.com/litelines #PyDataParis
tinyurl.com
alonsosilva/litelines-notebooks at main
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
020
Alonso Silva @alonsosilva.bsky.social · 29/09/2025
Featured in marimo newsletter 🤩 marimo.io/blog/newslet...
010
Alonso Silva @alonsosilva.bsky.social · 27/09/2025
Doing a slightly better documentation than just the README.md alonsosilvaallende.github.io/litelines/ge... Feeback?
041
Alonso Silva @alonsosilva.bsky.social · 16/09/2025
Want to visualize the response format constraints on the LLM when working in a Jupyter notebook? Then you might be interested in my new project `litelines`. Litelines lets you visualize the selected path by the LLM. It supports a Pydantic schema as a response format, as well as regular expressions.
194
Alonso Silva @alonsosilva.bsky.social · 28/07/2025
The recording of my presentation "Certainty-Guided Reasoning: A Dynamic Thinking Budget Approach" at the Laboratory for Information, Networking and Communication Sciences (LINCS) is now available: www.youtube.com/watch?v=8a44...
youtube.com
Certainty-Guided Reasoning in Large Language Models: A Dynamic Thinking Budget Approach
YouTube video by Pupusse LINCS
020
Alonso Silva @alonsosilva.bsky.social · 23/07/2025
New blog post: Constrain a language model not to use the letter 'e' alonsosilvaallende.github.io/blog/posts/2... In this post, I constrain a small language model (0.6B parameters) with a logits processor to accomplish something GPT-4o fails to do (see chatgpt.com/share/687908...).
110
Alonso Silva @alonsosilva.bsky.social · 18/07/2025
TIL how to add notebook cells running on @pyodide.org to Quarto projects, such as my blog: alonsosilvaallende.github.io/til/posts/20... It's very easy to change the Pyodide version if needed. Thanks @coatless.bsky.social for this great Quarto extension
032
Reposted by Alonso Silva
hailey @hailey.at · 18/07/2025
tfw u need neo4j to put together the polycule chart
4774
Alonso Silva @alonsosilva.bsky.social · 16/07/2025
New blog post: Understanding Logits Processors alonsosilvaallende.github.io/blog/posts/2... I start with a basic min length example, then force the model to check its answer, followed by making reasoning models stop thinking once they reach a thinking budget & finally force the model to think longer
000
Alonso Silva @alonsosilva.bsky.social · 11/07/2025
New blog post: Understanding Structured Outputs alonsosilvaallende.github.io/blog/posts/2... This is the power behind structured ouputs libraries such as Instructor and Marvin. I provide a basic example of extraction, a slightly more complex one, then classification, and finally SO in WebAssembly.
000
Alonso Silva @alonsosilva.bsky.social · 05/07/2025
New blog post: Understanding Function Calling I provide a basic example of FC, then a slightly more complex example by allowing an LM to use Python. I explain the conversational response as a tool trick. Finally, FC in the browser by using WebAssembly alonsosilvaallende.github.io/blog/posts/2...
000
Alonso Silva @alonsosilva.bsky.social · 30/06/2025
So I appear in the Celebration of 100 years of Bell Labs video with our LLM robots (a.k.a. IndustrialGPT) for less than a second :-D www.youtube.com/watch?v=Fu_I...
000
Alonso Silva @alonsosilva.bsky.social · 28/06/2025
New blog post: Understanding LLM Memory alonsosilvaallende.github.io/blog/posts/2... Using the Marimo extension for Quarto.
000
Alonso Silva @alonsosilva.bsky.social · 20/06/2025
New post: Understanding Chat Templates alonsosilvaallende.github.io/blog/posts/2...
alonsosilvaallende.github.io
Understanding Chat Templates – Homepage
000
Alonso Silva @alonsosilva.bsky.social · 17/06/2025
New post: Understanding Tokenizers alonsosilvaallende.github.io/blog/posts/2...
alonsosilvaallende.github.io
Understanding Tokenizers – Homepage
010
Alonso Silva @alonsosilva.bsky.social · 17/06/2025
Celebrating Bell Labs’ 100 years anniversary #NokiaBellLabs100
010
Alonso Silva @alonsosilva.bsky.social · 16/06/2025
TIL to do quizzes with JupyterQuiz: alonsosilvaallende.github.io/til/posts/20... I like this simple way to provide feedback to students. I’m using these quizzes to motivate my kid to learn new things.
010
Alonso Silva @alonsosilva.bsky.social · 15/06/2025
TIL you can embed a live REPL on a website: alonsosilvaallende.github.io/til/posts/20... I think this is a great tool for writing tutorials and blog posts.
alonsosilvaallende.github.io
Embed live REPL on a website – Homepage
010
Alonso Silva @alonsosilva.bsky.social · 20/05/2025
Finally meeting @remilouf.bsky.social !
010
Reposted by Alonso Silva
Cameron @cameron.stream · 29/04/2025
this was good
021
Alonso Silva @alonsosilva.bsky.social · 27/04/2025
The new release of logits-processor-zoo supports batch inference of the logits processors with vLLM `pip install logits-processor-zoo` Here is the repo: github.com/NVIDIA/logit... Here is the commit 😊: github.com/NVIDIA/logit...
000
Alonso Silva @alonsosilva.bsky.social · 23/04/2025
My talk "Building Knowledge Graph-Based Agents with Structured Text Generation" at PyData Global 2024 is now available on YouTube: www.youtube.com/watch?v=94yu... #PyData #PyDataGlobal @dottxtai.bsky.social @pydata.bsky.social
youtube.com
Alonso Silva - Building Knowledge Graph-Based Agents with Structured Text Generation
YouTube video by PyData
041
Alonso Silva @alonsosilva.bsky.social · 23/04/2025
More coverage of our work at @ieeespectrum.bsky.social 🤩
A snippet of the article that reads:
"The lab is also dipping into AI. One team is working on integrating large language models with robots for industrial applications. These robots have access to a digital twin model of the space they are in and have a semantic representation of certain objects in their surroundings. In a demo, a robot was verbally asked to identify missing boxes in a rack, and it successfully pointed out which box wasn’t found in its intended place, and when prompted travelled to the storage area and identified the replacement. The key is to build robots that can “reason about the physical world,” says Matthew Andrews, a researcher in the AI lab. A test system will be deployed in a warehouse in the United Arab Emirates in the next six months."
010
Alonso Silva @alonsosilva.bsky.social · 20/04/2025
Chat Templates Chat is an abstraction. LLMs know nothing about system messages, user messages, etc. The translation between messages and the received string representation is handled by the chat template. Code: github.com/alonsosilvaa... To try it: uv run raw.githubusercontent.com/alonsosilvaa...
000
Alonso Silva @alonsosilva.bsky.social · 20/04/2025
Next token The code can be found here: github.com/alonsosilvaa... If you have uv installed, you can try it with this command: uv run raw.githubusercontent.com/alonsosilvaa...
An app showing a user prompt "Hi" and the LLM probabilities to the reply word.
000
Reposted by Alonso Silva
Simon Willison @simonwillison.net · 18/03/2025
I figured out a minimal pattern for building a completely custom website using GitHub Actions and deploying the result to GitHub Pages, so I wrote it up as a TIL til.simonwillison.net/github-actio...
til.simonwillison.net
Building and deploying a custom site using GitHub Actions and GitHub Pages
I figured out a minimal pattern for building a completely custom website using GitHub Actions and deploying the result to GitHub Pages.
51438
Alonso Silva @alonsosilva.bsky.social · 16/02/2025
The answer is wrong but DeepSeek's reasoning traces are always fun to watch :-)
<think>
Okay, I need to help the user by providing timestamps and chapters for their given text. The text is a transcript of someone explaining how they set up a vector database using Milvus or Qdrant with song lyrics. They mentioned creating a table, inserting data, and querying it.
First, I'll read through the entire text to understand the flow. It starts by talking about two types of queries: one about a song with nightmares (Inter Sandman) and another about a guy with wounds (Stairway to Heaven). Then they mention that the results aren't perfect but still good.

I should break this down into logical chapters. Maybe start with an introduction, then setting up the database, inserting data, querying for the first song, querying for the second song, discussing imperfections, and a conclusion.

Looking at the timestamps provided earlier, each chapter is about 30 seconds to a minute long. I'll need to assign timestamps based on where each section starts in the original text.

I should make sure the chapters are clear and cover all main points without overlapping too much. Also, the timestamps should be accurate based on the user's initial example.

Finally, I'll present this in a structured way, using bold for chapter titles and italics for descriptions, along with the start and end times.
</think>
Here’s how you can structure the content into chapters with their respective timestamps:

Chapter 1: Introduction to Vector Databases
Timestamp: 00:00:00 - 00:01:00
Introduction to vector databases like Milvus or Qdrant and their relevance in handling text data.
020
Reposted by Alonso Silva
iadevs.bsky.social @iadevs.bsky.social · 18/01/2025
Revive la charla "Una introducción amigable a modelos de difusión" de Cristian Garcia disponible en YouTube: www.youtube.com/live/uHvP5uR... cc: @aastroza.bsky.social @sebastiandres.bsky.social @alonsosilva.bsky.social
youtube.com
IADevs 2024: Una introducción amigable a modelos de difusión, con Cristian Garcia de Google DeepMind
YouTube video by iadevs
061
Alonso Silva @alonsosilva.bsky.social · 16/01/2025
I use Kùzu but I must admit these are cute 😜
030
Reposted by Alonso Silva
iadevs.bsky.social @iadevs.bsky.social · 16/01/2025
Este jueves 23 de enero, @capetorch.bsky.social Senior MLE en @weightsbiases.bsky.social compartirá su experiencia en implementación de LLMs en producción junto a la comunidad IADevs en las oficinas de @fintual.bsky.social. Regístrate aquí: lu.ma/bcny2f4l Cupos limitados. ¡No te quedes fuera!
Encuentro de Verano el 23 de Enero de 19:00 a 21:00 hrs con Thomas Capelle de Weights and Biases.
052
Alonso Silva @alonsosilva.bsky.social · 14/01/2025
AI Eats the World www.youtube.com/watch?v=LGDa...
youtube.com
Benedict Evans: AI Eats the World | Slush 2024
YouTube video by Slush
010
Alonso Silva @alonsosilva.bsky.social · 12/01/2025
Air Katakana @airkatakana "¡ do not know javascript, i do not know how the internet works, and i have no idea how to make an app" "so how are you building all this?" "through my ability to tell claude what i want in a way it understands, and my ability to evaluate whether or not its output makes sense" 4:07 AM • 12/31/24 • 495K Views
020
Alonso Silva @alonsosilva.bsky.social · 01/01/2025
My 2025 resolution is to take more risks
It reads: Everything’s high risk if you’re a p*ssy
020
Alonso Silva @alonsosilva.bsky.social · 30/12/2024
Me entrevistaron para el podcast IA para los negocios de EvoAcademy ¡Espero que les guste! youtu.be/-56EfNIGy0k
youtu.be
Cómo evitar que la IA se equivoque - GraphRAG y Structured Output
YouTube video by EvoAcademy
000
Reposted by Alonso Silva
maartenbreddels.bsky.social @maartenbreddels.bsky.social · 17/12/2024
This is super cool, run an AI chatbot, pure Python fully in the browser using Solara. Great work @alonsosilva.bsky.social ❤️
021
Alonso Silva @alonsosilva.bsky.social · 16/12/2024
The pictures from the @iadevs.bsky.social conference are now available: photos.google.com/share/AF1Qip... It was a super cool conference!
Picture of the speaker presenting.Picture of the speaker presenting.Picture of the speaker presenting.Picture of Sebastian Flores, Alonso Astroza, and Alonso Silva (from left to right).
021