Duarte
@duarteocarmo.com
Born in 🇵🇹, based in 🇩🇰. Machine Learning Engineer (or whatever it's called now)
duarteocarmo.com
duarteocarmo.com/photos
github.com/duarteocarmo/
duarteocarmo.com
duarteocarmo.com/photos
github.com/duarteocarmo/
You use agents. I use agents. We all use agents.
Here's how I police them.
duarteocarmo.com/blog/how-to...
Here's how I police them.
duarteocarmo.com/blog/how-to...
How to police your agents
We're building more than ever. But how do we manage the complexity? We police our agents.
duarteocarmo.com
January 23, 2026 at 3:30 PM
You use agents. I use agents. We all use agents.
Here's how I police them.
duarteocarmo.com/blog/how-to...
Here's how I police them.
duarteocarmo.com/blog/how-to...
opencode agent create
Go ahead, give it a shot.
Go ahead, give it a shot.
January 6, 2026 at 11:00 AM
opencode agent create
Go ahead, give it a shot.
Go ahead, give it a shot.
I don't like ads. You don't like ads. We don't like ads.
Limpa is like AdBlock + SponsorBlock but for podcasts.
GitHub repo below
open.substack.com/pub/duarteo...
Limpa is like AdBlock + SponsorBlock but for podcasts.
GitHub repo below
open.substack.com/pub/duarteo...
January 5, 2026 at 6:42 PM
I don't like ads. You don't like ads. We don't like ads.
Limpa is like AdBlock + SponsorBlock but for podcasts.
GitHub repo below
open.substack.com/pub/duarteo...
Limpa is like AdBlock + SponsorBlock but for podcasts.
GitHub repo below
open.substack.com/pub/duarteo...
Running a web server + a worker used to be a pain.
Thanks Claude, didn't know about this one.
github.com/nickstennin...
Thanks Claude, didn't know about this one.
github.com/nickstennin...
GitHub - nickstenning/honcho: Honcho: a python clone of Foreman. For managing Procfile-based applications.
Honcho: a python clone of Foreman. For managing Procfile-based applications. - nickstenning/honcho
github.com
January 4, 2026 at 3:00 PM
Running a web server + a worker used to be a pain.
Thanks Claude, didn't know about this one.
github.com/nickstennin...
Thanks Claude, didn't know about this one.
github.com/nickstennin...
Now with agents you can have super meaningful commit messages. It's incredible.
Lul
Lul
January 2, 2026 at 7:01 PM
Now with agents you can have super meaningful commit messages. It's incredible.
Lul
Lul
Django background tasks feedback.
Me like.
Me like.
January 1, 2026 at 4:00 PM
Django background tasks feedback.
Me like.
Me like.
complexipy
kreuzberg
throttled-py
batchata
I'm living under a rock? Never heard of any of these. Tasty though.
tryolabs.com/blog/top-py...
kreuzberg
throttled-py
batchata
I'm living under a rock? Never heard of any of these. Tasty though.
tryolabs.com/blog/top-py...
Top Python libraries of 2025
Explore our 11th annual Top Python Libraries roundup, featuring two curated Top 10 lists for General Use and AI / ML / Data tools that matter today.
tryolabs.com
January 1, 2026 at 11:12 AM
complexipy
kreuzberg
throttled-py
batchata
I'm living under a rock? Never heard of any of these. Tasty though.
tryolabs.com/blog/top-py...
kreuzberg
throttled-py
batchata
I'm living under a rock? Never heard of any of these. Tasty though.
tryolabs.com/blog/top-py...
Dirty little secret.
Cancel Claude code, cancel ChatGPT plus.
Github copilot is 10 bucks a month. And OpenCode can use them all
🤫
Cancel Claude code, cancel ChatGPT plus.
Github copilot is 10 bucks a month. And OpenCode can use them all
🤫
January 1, 2026 at 8:00 AM
Dirty little secret.
Cancel Claude code, cancel ChatGPT plus.
Github copilot is 10 bucks a month. And OpenCode can use them all
🤫
Cancel Claude code, cancel ChatGPT plus.
Github copilot is 10 bucks a month. And OpenCode can use them all
🤫
Github actions in a nutshell: Comment in an issue -> get a notification for a failed action 😂
December 30, 2025 at 12:56 PM
Github actions in a nutshell: Comment in an issue -> get a notification for a failed action 😂
2 themes built
1 attempt at building an lsp for beancount account completion
34 completion engines later.
Done. Now we can carry on.
github.com/duarteocarm...
1 attempt at building an lsp for beancount account completion
34 completion engines later.
Done. Now we can carry on.
github.com/duarteocarm...
dotfiles/.config/nvim-minimal/init.lua at master · duarteocarmo/dotfiles
Contribute to duarteocarmo/dotfiles development by creating an account on GitHub.
github.com
December 30, 2025 at 11:51 AM
2 themes built
1 attempt at building an lsp for beancount account completion
34 completion engines later.
Done. Now we can carry on.
github.com/duarteocarm...
1 attempt at building an lsp for beancount account completion
34 completion engines later.
Done. Now we can carry on.
github.com/duarteocarm...
Nothing like the fresh smell of a new Django project.
December 28, 2025 at 12:42 PM
Nothing like the fresh smell of a new Django project.
Part of my move away from LazyVim.
My config is now 200 lines of pure lua and I love it.
github.com/duarteocarm...
My config is now 200 lines of pure lua and I love it.
github.com/duarteocarm...
GitHub - duarteocarmo/cursor-themes: Cursor inspired themes for neovim
Cursor inspired themes for neovim . Contribute to duarteocarmo/cursor-themes development by creating an account on GitHub.
github.com
December 26, 2025 at 9:06 PM
Part of my move away from LazyVim.
My config is now 200 lines of pure lua and I love it.
github.com/duarteocarm...
My config is now 200 lines of pure lua and I love it.
github.com/duarteocarm...
This month in Retrospectiva:
- Neovim
- OpenCode
- A kitchen gadget that changed my life
- A new health tracker I'm in love with
- And some music.
Last post of 2025! Enjoy the holidays :)
open.substack.com/pub/duarteo...
- Neovim
- OpenCode
- A kitchen gadget that changed my life
- A new health tracker I'm in love with
- And some music.
Last post of 2025! Enjoy the holidays :)
open.substack.com/pub/duarteo...
December 26, 2025 at 1:30 PM
This month in Retrospectiva:
- Neovim
- OpenCode
- A kitchen gadget that changed my life
- A new health tracker I'm in love with
- And some music.
Last post of 2025! Enjoy the holidays :)
open.substack.com/pub/duarteo...
- Neovim
- OpenCode
- A kitchen gadget that changed my life
- A new health tracker I'm in love with
- And some music.
Last post of 2025! Enjoy the holidays :)
open.substack.com/pub/duarteo...
We're definitely fighting the fight. The amount of trash out there hasn't stopped increasing.
As always, the less code the better.
www.youtube.com/watch?v=Ioi...
As always, the less code the better.
www.youtube.com/watch?v=Ioi...
The War on Slop – swyx
Why we need to eliminate low-quality code and work in AI engineering.Speaker: swyx | Organizer, AI Engineerhttps://x.com/swyxhttps://www.linkedin.com/in/sh...
www.youtube.com
December 25, 2025 at 4:00 PM
We're definitely fighting the fight. The amount of trash out there hasn't stopped increasing.
As always, the less code the better.
www.youtube.com/watch?v=Ioi...
As always, the less code the better.
www.youtube.com/watch?v=Ioi...
I have to say I'm a bit sick of LazyVim. I'm moving to NeoVim's default package manager: Pack.
Fast, simple.
Not finished, but here it is: github.com/duarteocarm...
Fast, simple.
Not finished, but here it is: github.com/duarteocarm...
dotfiles/.config/nvim-minimal/init.lua at master · duarteocarmo/dotfiles
Contribute to duarteocarmo/dotfiles development by creating an account on GitHub.
github.com
December 24, 2025 at 8:22 PM
I have to say I'm a bit sick of LazyVim. I'm moving to NeoVim's default package manager: Pack.
Fast, simple.
Not finished, but here it is: github.com/duarteocarm...
Fast, simple.
Not finished, but here it is: github.com/duarteocarm...
In this post I dive deep into using LLMs for nutritional content analysis of meal descriptions.
We talk DSPy, GEPA, on-device inference, and more!
duarteocarmo.com/blog/from-n...
We talk DSPy, GEPA, on-device inference, and more!
duarteocarmo.com/blog/from-n...
From NutriBench to Taralli: How far can you take a prompt?
Using NutriBench to optimize Taralli, my nutritional estimation iOS app.
duarteocarmo.com
December 23, 2025 at 6:33 PM
In this post I dive deep into using LLMs for nutritional content analysis of meal descriptions.
We talk DSPy, GEPA, on-device inference, and more!
duarteocarmo.com/blog/from-n...
We talk DSPy, GEPA, on-device inference, and more!
duarteocarmo.com/blog/from-n...
Publishing an app in the app store is such a broken process.
- Builds, versions
- App store reviews (sometimes 30 mins, sometimes 4 days)
- Publish to app store connect takes 1 hour
Is there a (SIMPLE) CI/CD, push, increment app version publish (SwiftUI)
- Builds, versions
- App store reviews (sometimes 30 mins, sometimes 4 days)
- Publish to app store connect takes 1 hour
Is there a (SIMPLE) CI/CD, push, increment app version publish (SwiftUI)
December 22, 2025 at 11:00 AM
Publishing an app in the app store is such a broken process.
- Builds, versions
- App store reviews (sometimes 30 mins, sometimes 4 days)
- Publish to app store connect takes 1 hour
Is there a (SIMPLE) CI/CD, push, increment app version publish (SwiftUI)
- Builds, versions
- App store reviews (sometimes 30 mins, sometimes 4 days)
- Publish to app store connect takes 1 hour
Is there a (SIMPLE) CI/CD, push, increment app version publish (SwiftUI)
thinking
no-thinking
reasoning_effort
max_reasoning_tokens
thinking_enabled
include_reasoning
reasoning_summaries
It's just me? Or are y'all also a bit sick of this 😂
no-thinking
reasoning_effort
max_reasoning_tokens
thinking_enabled
include_reasoning
reasoning_summaries
It's just me? Or are y'all also a bit sick of this 😂
December 19, 2025 at 4:00 PM
thinking
no-thinking
reasoning_effort
max_reasoning_tokens
thinking_enabled
include_reasoning
reasoning_summaries
It's just me? Or are y'all also a bit sick of this 😂
no-thinking
reasoning_effort
max_reasoning_tokens
thinking_enabled
include_reasoning
reasoning_summaries
It's just me? Or are y'all also a bit sick of this 😂
Do I have to come here everytime a new model comes out?
December 19, 2025 at 11:00 AM
Do I have to come here everytime a new model comes out?
When do they send the badge? @OpenRouterAI
December 18, 2025 at 8:43 PM
When do they send the badge? @OpenRouterAI
A good rule of thumb.
Have an eval script ready for your application.
New model comes out? Add the string and run.
Will answer the "Should we be using this model?" real quick.
Measure:
- Time to first token
- Latency
- Verbosity
- Quality (llm as judge)
Have an eval script ready for your application.
New model comes out? Add the string and run.
Will answer the "Should we be using this model?" real quick.
Measure:
- Time to first token
- Latency
- Verbosity
- Quality (llm as judge)
December 18, 2025 at 4:00 PM
A good rule of thumb.
Have an eval script ready for your application.
New model comes out? Add the string and run.
Will answer the "Should we be using this model?" real quick.
Measure:
- Time to first token
- Latency
- Verbosity
- Quality (llm as judge)
Have an eval script ready for your application.
New model comes out? Add the string and run.
Will answer the "Should we be using this model?" real quick.
Measure:
- Time to first token
- Latency
- Verbosity
- Quality (llm as judge)
Alright @opencode I'm sold.
December 18, 2025 at 3:08 PM
Alright @opencode I'm sold.
Pro tip: You have multiple lambda intertwined, it's a mess.
Direct them all into the same log group and run:
aws logs tail /aws/lambda/yourlambdaname --follow
Now everything's in one place
Direct them all into the same log group and run:
aws logs tail /aws/lambda/yourlambdaname --follow
Now everything's in one place
December 18, 2025 at 11:00 AM
Pro tip: You have multiple lambda intertwined, it's a mess.
Direct them all into the same log group and run:
aws logs tail /aws/lambda/yourlambdaname --follow
Now everything's in one place
Direct them all into the same log group and run:
aws logs tail /aws/lambda/yourlambdaname --follow
Now everything's in one place