GitHub Profile Data (fetched 2026-03-05)

Profile

• Username: ruanchaves (Ruan Chaves)
• Bio: Senior AI Engineer with 5+ years of experience delivering real-world solutions using Generative AI, LLMs, and NLP.
• Location: Brazil
• Website: https://ruanchaves.github.io/
• LinkedIn: in/ruanchaves
• Email: ruanchaves93@gmail.com
• Followers: 62
• Following: 30
• Repositories: 75
• Stars given: 67

Achievements

• Pull Shark (x3)
• YOLO
• Quickdraw
• Starstruck
• Arctic Code Vault Contributor

Pinned Repositories

hashformers (77 stars, 5 forks)

• Description: Accurate word segmentation for hashtags and text, powered by Transformers and Beam Search. A scalable alternative to heuristic splitters and massive LLMs.
• Language: Python
• License: MIT
• Created: 2020-05-21
• Last updated: 2026-02-21
• Topics: deep-learning, hashtag-segmentor, large-language-models, llms, natural-language-processing, nlp, paper, segmentation, sentiment-analysis, sentiment-classification, sentiment-polarity, spacy, spacy-extension, spacy-extensions, transformers, tweet-analysis, tweets-classification, twitter, twitter-sentiment-analysis, word-segmentation
• Notable: Recognized as state-of-the-art at LREC 2022 in the paper “HashSet - A Dataset For Hashtag Segmentation” by researchers from IIT. Leverages GPT-2 and beam search for accurate, multilingual hashtag and text segmentation.

napolab (72 stars, 3 forks)

• Description: The Natural Portuguese Language Benchmark (Napolab). Stay up to date with the latest advancements in Portuguese language models and their performance across carefully curated Portuguese language tasks.
• Language: Python
• License: MIT
• Created: 2023-03-29
• Last updated: 2026-01-26
• Topics: benchmarks, catalan, datasets, english, galician, hate-speech, huggingface, huggingface-transformers, large-language-models, nlp, portuguese, python, question-answering, semantic-similarity, spanish, text-simplification, textual-entailment, transformers
• Notable: Key finding - the performance gap between general-purpose LLMs and Portuguese-specific models is smaller than previously believed. Exposes systemic issues in LLM benchmarking (investigation gap, data contamination).
• Leaderboard: https://huggingface.co/spaces/ruanchaves/napolab
• Master Thesis: https://www.um.edu.mt/library/oar/handle/123456789/120557

elmo (11 stars, 2 forks)

• Description: Supporting code for the paper “Portuguese Language Models and Word Embeddings: Evaluating on Semantic Similarity Tasks”.
• Language: Jupyter Notebook
• Topics: elmo, embeddings, natural-language-processing, natural-language-understanding, nlp, portuguese, portuguese-language, semantic-similarity, textual-entailment

song2vec (6 stars, 0 forks)

• Description: Telegram bot that recommends songs as YouTube playlists through gensim’s word2vec
• Language: Python

assin (5 stars, 3 forks)

• Description: Supporting code for the paper “Multilingual Transformer Ensembles for Portuguese Natural Language Tasks”.
• Language: Jupyter Notebook
• Topics: bert, natural-language-processing, natural-language-understanding, nlp, portuguese, portuguese-language, roberta, semantic-similarity, textual-entailment, transformers

reddit-html-archiver-image-plugin (5 stars, 0 forks)

• Description: image downloader plugin for reddit-html-archiver
• Language: Python

BERT-WS (0 stars, 0 forks)

• Description: Supporting code for the paper “Domain Adaptation of Transformers for English Word Segmentation”.
• Language: Python

Other Notable Repos

Zero-Shot-Entity-Linking (2 stars)

• Zero-shot Entity Linking with blitz start in 3 minutes.

medical-assistant-bot

• A medical question-answering system that can effectively answer user queries related to medical diseases.
• Language: Jupyter Notebook

countgpt

• Language: Python

qa-dataset

• Language: Python

ml-tech-assessment

• Language: Python

Open Source Contributions (from github.md)

argilla-io/argilla

• Fixed bugs and shipped features related to semi-supervised learning (SSL) during internship at Argilla.
• Argilla was acquired by Hugging Face in June 2024 (~$10M deal).

huggingface/transformers (PR #10823)

• Modified the Trainer class for simultaneous Ray Tune and Weights & Biases execution.

nathanshartmann/portuguese_word_embeddings (PR #11)

• Fixed a severe bug in the evaluation procedure. Documented in research paper.

facebookresearch/BLINK (PR #25)

• Fixed a parameter bug in the script for the BLINK benchmark.

awslabs/mlm-scoring (PR #12)

• Addressed an installation instruction issue for the mlm-scoring library.

All Repos (75 total, sorted by stars)