Bolsista Mestre – Cientista de Dados, NLP, Sistemas de Busca Semântica

This is a fully remote position, open to applicants in Brazil.

📋 Description

• Text preprocessing and enrichment: Cleaning, tokenization, lemmatization, and noise removal in textual documents

• Generation and management of embeddings: Creating document and query embeddings using models such as Sentence-BERT, OpenAI ada, or similar

• Development of semantic search pipelines, information retrieval, and applications with RAG: Building pipelines that combine: query → embedding → vector search → re-ranking (optional)

• Assessing the quality of retrieval and generated responses in NLP and RAG scenarios, using metrics like recall@k and MRR

• Adaptation and experimentation with models in PyTorch or TensorFlow

• Documentation and versioning: Documenting pipelines, technical decisions, and experimental results

⛳️ Requirements

• Ongoing or completed Master's degree

• Computer Science, Computer/Software Engineering, Information Systems, Statistics, Applied Mathematics, Electrical Engineering, Data Science, or related fields in Exact Sciences and Engineering

• Proficiency in Python programming with a strong understanding of typing, data manipulation (pandas/polars), and use of virtual environments

• Practical experience with NLP, including: tokenization, stemming/lemmatization, stopword removal, and vectorization (TF IDF, word2vec, or embeddings)

• Applied knowledge of embeddings, semantic search, and RAG techniques (e.g., Sentence BERT, current embedding models or similar), as well as familiarity with vector databases (e.g., FAISS, ChromaDB, Pinecone, Qdrant)

• Experience with at least one deep learning framework (PyTorch or TensorFlow)

• Understanding of analytical pipelines and experimentation in data science applied to NLP (e.g., extraction → preprocessing → embedding → search/classification → generation with context) and versioning with Git

🏝️ Benefits

• Scholarship: R$ 9,000.00