Thiago Marques
Data Scientist
Manaus, Amazonas, Brazil
Summary
👤
SeniorThiago Marques is a Senior Data Scientist with 11 years of experience building production ML and NLP solutions, currently developing GenAI conversational agents and RAG pipelines at Stone. He combines academic depth—a M.Sc. in Computer Science focused on deep learning for duplicate bug detection—with hands-on engineering across LLMs, speech models, vector DBs and scalable infra (FastAPI, Docker, Spark, GCP/AWS). At Instituto de Pesquisas Eldorado he led novel multi-modal systems (STT/TTS, XML PAI generation, advanced RAG and LLM agents) and automated SKU grouping via textual similarity, demonstrating a knack for turning research ideas into MVPs. Past roles span anomaly detection, recommendation and routing optimization work for enterprise clients, and earlier full-stack/Android engineering that informs his product-aware approach. He is based in Santa Catarina, Brazil, and is comfortable bridging research, applied ML and deployment—often combining open-source models and engineering pragmatism to deliver business impact.
11 years of coding experience
Github Skills (27)
notebook10
javascript10
python10
classification10
django10
attention-mechanism10
machine-learning10
java10
smile10
sklearn10
android10
supervised-learning10
scikit-learn10
percentage10
jupyter-notebook10
Programming languages (4)
Github contributions (5)
thiagomarquesrocha/siameseQAT
Feb 2019 - Aug 2021
SiameseQAT, a duplicate bug report detection method that considers not only information on individual bugs, but also collective information from bug clusters. SiameseQAT combines attention mechanisms, which were not previously used in this task, with a novel loss function called Quintet Loss, that considers the centroid of duplicate bug report representation clusters andtheir contextual information.
Contributions:7 releases, 333 commits, 2 PRs in 2 years 6 months
attention-mechanism
thiagomarquesrocha/Kaio-machine-learning-human-face-detection
Dec 2016 - Oct 2017
Machine Learning project a case study focused on the interaction with digital characters, using a character called "Kaio", which, based on the automatic detection of facial expressions and classification of emotions, interacts with humans by classifying emotions and imitating expressions
Contributions:73 commits, 64 pushes, 1 branch in 10 months
classificationmachine-learningdetectionsklearnjavascript
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.