Duarte O.Carmo
The personal website of Duarte O.Carmo. A technologist/consultant from Lisbon, now based in Copenhagen.
- Posts
- 105 from 11 May 2016 to 8 Oct 2026
- Posting
- weekly (27 in the last year)
- Kind
- personal blog
- Feed
- https://duarteocarmo.com/feed.xml
2026 22
- Retrospectiva #12
- AMALIA vs. smaller reasoning models
- Retrospectiva #11
- How I use Pi
- Bagaço v3: Now with PDFs and Wikipedia!
- Retrospectiva #10
- Ginjinha: Pretraining LLMs on European Portuguese
- Retrospectiva #9
- How I Hermes
- Retrospectiva #8
- TTS doesn't suck anymore
- The largest open pretraining dataset for European Portuguese
- Retrospectiva #7
- AMÁLIA and the future of European Portuguese LLMs
- Retrospectiva #6
- An opinionated running dashboard
- Portuguese variety identification: The bitter lesson
- Retrospectiva #5
- Bagaço: A pretraining dataset for European Portuguese
- Retrospectiva #4
- How to police your agents
- Limpa: Ad-Free podcasts powered by LLMs
2025 18
- Retrospectiva #3
- From NutriBench to Taralli: How far can you take a prompt?
- Retrospectiva #2
- Book release: DeepSeek in Practice
- TTS still sucks
- Retrospectiva #1
- Faísca: The modern LLM stack in a single script
- AI: the utility bubble - for Publico.pt
- Drowning in News
- Common misconceptions about AI
- A Benchmark for language models on European Portuguese
- MCPs are mostly hype
- Evals are all you need
- Taralli: Home-Cooked Food Tracking Without the BS
- Goodbye Kindle, I don't think I'll miss you.
- You also hate SQL? Let the LLM handle it
- Simple AI tools
- What the hell Is GPQA, anyway?
2024 16
- RAG tricks from the trenches
- An ode to the Marathon
- Upgrading this website's podcast with F5-TTS
- What if ChatGPT's memory was a knowledge graph?
- Classification in the age of LLMs: The emoji problem
- Around Iceland in 6 days
- Tanda Runner: A personalized running dashboard
- Do things that don't scale
- How I self-host in 2024
- The best keyboard I've ever had
- An experiment with Gemma 2B and a Portuguese classic
- NewsHavn: Danish news, in English
- mistral-doc: Fine-tuning an LLM on my ChatGPT conversations
- What's in my NOW?
- Self-hosting my personal LLM (but not quite)
- Rebuilding /photos
2023 12
- You can now listen to this blog
- Should you go into management consulting? Maybe.
- Changelog neural search
- A poor man's guide to fine-tuning Llama 2
- aicoverlettercreator.com
- Supercharging my Telegram group with the help of ChatGPT
- Fine-tuning FLAN-T5 to replace my friends
- Governo Sombra transcripts
- LLMs in production: lessons learned
- parlabot - ask the Portuguese parliament
- An opinionated Python boilerplate
- On Rust
2022 12
- Scaling Machine Learning microservices
- Monitoring Machine Learning APIs
- Goodbye Apple Watch
- infrequent.app - stay in touch with those who matter
- Monitoring ML models with FastAPI and Evidently AI
- 05-08-2022 Down from the Cloud
- 26-06-2022 A recipe recommendation system
- 10-06-2022 Serving models with FastAPI: It's not just about the speed
- 12-04-2022 Mac apps you didn't know you needed
- 21-03-2022 A simple system to stay in touch with people
- 10-02-2022 Visualizing every job in the world
- 06-01-2022 Simple software
2021 8
- 26-12-2021 NFTs are dumb. Let's make some
- 21-11-2021 How to work from home - revisited
- 06-10-2021 Four years of Python
- 20-09-2021 Self-hosting my Instagram profile
- 24-06-2021 The short tale of an online scam
- 24-05-2021 Run every day
- 11-04-2021 How to build a newsletter using Python and FastAPI
- 15-01-2021 Hacking on my finances (Part 2: Beancount on Beanstalk)
2020 9
- 23-11-2020 Building a Telegram bot in Python to track your portfolio
- 18-10-2020 Hacking on my finances
- 23-07-2020 Can I go home yet?
- 22-06-2020 NbHub: Share your jupyter notebooks with the world
- 18-06-2020 Studying in Denmark as an international student
- 05-06-2020 VIM for Python development (and not only)
- 22-05-2020 Neural Networks for Linear Regressions using Python
- 25-03-2020 How to work from home
- 25-02-2020 Managing kindle highlights with Python and GitHub