Artificial Intelligence
[Open soon]
Work description
For the first objective, the student must experiment with generative models for summary generation using prompt engineering. The second objective encompasses the definition of the architecture, which should include a scheduler responsible for updating current reports, an orchestrator agent, a narrative identifier agent, a recommender agent, agents specialized in summarization, a summary aggregator agent, and a component for accessing a relational database. For the third objective, the student will need to implement the architecture defined in the second objective. The implementation should consider a framework specialized in agentic pipelines, such as LangGraph. Finally, for the fourth objective, the student will integrate pre-trained models into the narrative extraction and recommendation agents. Several experiments will be necessary in the system to identify the contribution of each agent and component. Therefore, ablation experiments will be performed. Finally, experiments evaluating different LLM models for summarization and aggregation methods will be executed.
Academic Qualifications
Bachelor's degree in computer science, Artificial Intelligence, Data Science, or equivalent.
Minimum profile required
Proven knowledge and experience in LLMs/NLP such as fine-tuning and prompting.Fluent in Portuguese. Advanced knowledge of English (primarily reading and writing).
Preference factors
Some experience with natural language processing, classification methods, agentic AI, and generative models. Previous participation in research projects. Writing and publishing scientific articles.
Application Period
Since 25 Jun 2026 to 09 Jul 2026
[Open soon]
Centre
Artificial Intelligence and Decision Support