From failing to study medicine ➡️ BSc industrial engineer ➡️ MSc computer scientist.
Life can be strange, so better enjoy it.
I´m sure I do by: 👨🏽‍🍳 Cooking, 👨🏽‍💻 Coding, 🏆 Committing.

• Synthetic Data - Weaviate Podcast #118! - podcast
• SmolAgents - From Bells and Whistles to Agents and Tools - slides video
• No data? No problem! - synthetic data to the rescue - slides video
• Practical AI Podcast - Towards high-quality (maybe synthetic) datasets - podcast
• Code Together Podcast Intel Software - Scaling LLM Datasets with Less Effort Using Argilla - video
• Mastering LLMs - Creating, curating, and cleaning data for LLMs - slides video
• 🧼 From GPU-poor to data-rich - data quality practices for LLM fine-tuning - slides
• Deeplearning.ai LLM workshop - get started with Argilla for human- and distilabel for AI feedback - video
• NLP Healthcare Summit 2023 - Smart Shortcuts for Bootstrapping a Healthcare NER Project - video
• Anyscale Ray Europe Meetup - Smart shortcuts for Bootstrapping a Text Classification project - video

• useful code snippets
• short videos

• Giskard 🐢 (2025-present) - Secure your LLM Agents.
• Pruna 🍑 (2025-present) - Make your AI models cheaper/faster/smaller/greener!
• Hugging Face 🤗 (2024-2025) - The AI community building the future.
• Argilla (2022-2025) - data annotation and monitoring for enterprise NLP.
• Pandora Intelligence (2020-2022) - an independent intelligence company, specialised in security risks.

• awesome-ai-guardrails - A collection of awesome AI guardrail resources
• observers - A Lightweight Library for AI Observability
• dataset-viber - Data viber is your chill repo for data collection and vibe checks
• concise-concepts - a word similarity approach to few-shot NER
• fast-sentence-transformers - simply, faster, sentence-transformers
• classy-classification - a quick and dirty few-shot text classification solution
• crosslingual-coreference - a multi-lingual CoRef resolver using cross-lingual training
• adept-augmentations - a Python library aimed at dissecting and augmenting NER training data
• spacy-setfit - a Python library aimed at facilitating easy SetFit usage in spaCy

• Haystack - small feature and CI/CD updates
  • InMemoryDatabase - Serialization + to and from disk methods
  • GitHub Actions - caching for pip environment
• spaCy - several additions to the spacy-universe
  • spanmarker - added .pipe() method to spaCy integration
  • spacy-dbpedia-spotlight - added a batch processing functionality
  • spacy-fishing - added a batch processing functionality + bug fixes
  • spacy-opentapioca - added a batch processing functionality
• streamlit-url-fragment - resolved Python versioning issues
• allennlp-models - added a batch processing functionality
• mutate - resolved Python versioning issues and added PyPI support
• rebel - added a batch processing functionality
• trl - updated RLHF documentation for PPOTrainer
• vicinity
  • added vector search on serializable item types
  • improved docs
• semhash
  • added deduplication based on entropy score
  • refactor API design data model

• Bonfari - small to medium sustainable scale projects in Gambia 🇬🇲
• 510 red-cross - occasional projects to improve humanitarian aid with data