BrenoAV

About Me

I have over 4 years of experience in the field of Artificial Intelligence, combining academic research with industry projects in areas such as computer vision, natural language processing, and time series forecasting.

2024–Present — Master’s Degree Student (UFCG)
Research Focus: Retrieval-Augmented Generation (RAG) for QA systems applied to Brazilian Grid Network Procedures
2022–2024 — Researcher at Ford Motor Company
Worked on projects involving computer vision (object detection), time series forecasting using statistical and machine learning models, and natural language processing (semantic search).
2021–2022 — Scientific Initiation Scholar (PIBIC/CNPq – UFCG)
Conducted research on computer vision, specifically in object classification tasks.
2016–2022 — B.Sc. in Electrical Engineering (UFCG)
Final Paper: Deep Learning for Industrial Surface Defect Classification

Outside of my professional and academic work, I enjoy running, watching TV shows and movies, listening to podcasts, and exploring astronomy content for fun.

Publications

VIEIRA, BIANCA ALENCAR; VIEIRA, BRENO ALENCAR. Sustainable Potential of Renewable Energies in the Brazilian Northeast . RGSA (ANPAD), v. 18, p. e010554, 2024.

Portfolio

Agentic Job Application Assistant

A smart assistant that obtains information from the natural language user input using the power of LLM tooling.

Technologies used: Python, Litellm, AI Agent

View on GitHub

Events

Third Place in DatathONS 2025

Participated in Hackathon provided by the National Electric System Operator (ONS) 2025 (Sep 27 – Sep 28, 2025), which focused on how to make the open data portal (https://dados.ons.org.br/) more accessible and easy to use with Natural Language. Our solution was HorizONS, an AI Agent that connects ONS data to new horizons of interpretation, democratizing access for managers, analysts, and researchers. The solution integrates an AI Agent with RAG, querying PostgreSQL and pgvector databases, integrating with extra documents and data dictionaries, and featuring additional skills like dynamic internet search and arithmetic operations for advanced analyses and automatic aggregations (e.g., calculation of total curtailment by source).

Technologies used: n8n, Python, Docker, Postgres

Presentation at SIMPASE XIV 2025

Presented the work “Sistema Baseado em RAG para Consulta Inteligente aos Procedimentos de Rede do ONS” (RAG-Based System for Intelligent Query of ONS Network Procedures) at the XIV SIMPÓSIO INTERNACIONAL DE ENGENHARIA ELÉTRICA E DE SISTEMAS (SIMPASE XIV) in Foz do Iguaçu, PR, from August 4 to 6, 2025. The event, promoted by CIGRE-Brasil, is the main forum on automation of electric power systems, focusing on themes like Artificial Intelligence and digital transformation in the sector.

First Place in Hackathon CREA-PB 2025

Participated in Hackathon CREA-PB 2025 (May 30 – June 1, 2025), focused on transforming Crea-PB services using Artificial Intelligence. ART (Technical Responsibility Annotation) is a formal document that engineers fill out to officially register their responsibility for a project or service. As a team, we developed an AI tool to automate the completion of the ART, streamlining the document’s generation and validation steps.

Technologies used: LangChain, Streamlit

View Publication →