David Gasquez
@davidgasquez
Followers
2K
Following
12K
Media
249
Statuses
5K
Learning about anything and everything all of the time. Open source, open data, open protocols. Data stuff @protocollabs. Prev more data things @buffer.
Granada, Spain
Joined December 2009
RT @GergelyOrosz: This is a reminder that it’s possible to sponsor @simonw GitHub with $10/mo. He sends an additional monthly wrap-up to th….
0
10
0
RT @davidgasquez: Now making the agents in the tool chose the final name of the tool using the tool itself. 😜
0
1
0
Now making the agents in the tool chose the final name of the tool using the tool itself. 😜
Finally had some time to publish a simple vibecoded tool I (and Claude Code) built to explore PydanticAI, Arbitron. Replicates a contest where jurors evaluate candidates in pairwise comparisons that get turned into a leaderboard.
0
1
2
You initialize an arbitrary set of agents with contex, values, tools, . and then run a ranking algorithm (ELO, bradley terry, . ) to get a leaderboard given the agents' choices. It is fun to run arbitrary contest with random personas!.
0
0
0
Finally had some time to publish a simple vibecoded tool I (and Claude Code) built to explore PydanticAI, Arbitron. Replicates a contest where jurors evaluate candidates in pairwise comparisons that get turned into a leaderboard.
1
0
2
Wrote a post about some things I've been doing to make my projects more LLM friendly. Spoiler alert: it makes the projects more human friendly too! . Any other ideas or suggestions?
1
0
6
0
14
0
When people are pressured to meet a target value there are three ways they can proceed:. 1) They can work to improve the system.2) They can distort the system.3) Or they can distort the data. — Donald Wheeler.
0
0
9
Wrote a few thoughts on how we could improve experiments like Deepfunding after participating in some fun competitions. Excited to continue exploring how open/transparent ML competitions can be used to scale human judgement!.
2
1
19
RT @fishnets88: My take was a bit different, when I think back of the lessons that come out of that group it's these vibes:. - allow yourse….
0
3
0
Open Source Observer (@OSObserver) is an amazing data engineering project. Check out their repository if you want to learn even more!.
1/ There are an overwhelming number of options for data engineering today. Here's today's choices at @OSObserver :. For cloud infrastructure: @googlecloud .For data orchestration: @dagster .For data transformations: @SQLMesh .For data ingest: @dltHub .For data lake tables:.
0
3
9
RT @dwddao: Apply to our Impact Evaluator Research Retreat 2025 from July 26th to Aug 10th, 2025, Iceland. We are funding accommodation a….
0
7
0
RT @PyDataGRX: ¡No te pierdas el 11º MeetUp de PyData Granada el 20/02/2025! 🤖 Exploraremos Agentes IA y Pacientes Virtuales. Ven a Novatec….
0
3
0
A benefit of keeping a Personal Knowledge Base with lots of tips is that these files are excellent prompts for LLMs! . - - - -
0
2
9
I've found Cursor Agent to be a much more useful tool with these tricks. I've seen it do web searches to check documentation, use `uvx` to run CLI Python commands, and even follow random links in with `markitdown` to learn more about a topic.
0
0
6
Had a lot of fun teaching Cursor Agent new tricks. ✨ . - How to search the web from the CLI .- How to run any Python / JavaScript code without installing dependencies .- How to read links as Markdown
1
0
5
Wrote a post about debugging @dagster in VSCode / Cursor! . If you're like me and use a lot of notebooks to write code, this might be a great alternative that relies on VSCode / Cursor debugger for a smooth development experience with Dagster or similar libraries.
2
3
19