Joaquim Breno Brito Cavalcante
Professional Summary
Machine Learning Engineer with a strong background in deep learning, signal processing, and large-scale model deployment. Experienced in bridging research and product using reproducible pipelines and modern MLOps practices. Skilled in developing data-driven solutions for structured and unstructured data across GCP and AWS.
Education
Federal University of Paraíba (UFPB)
In ProgressM.Sc. in Computer Science • 📍 João Pessoa, Brazil
Focus: Machine Learning, Deep Learning, AI Research
Federal University of Paraíba (UFPB)
Graduated Oct 2025B.E. in Computer Engineering • 📍 João Pessoa, Brazil
Relevant coursework: Machine Learning, Deep Learning, Signal Processing, Algorithms
Experience
Machine Learning Researcher — Music AI
Sep 2025 – Present📍 Salt Lake City, UT
Develop ML models to auto tune and enhance vocals and data pipelines for large-scale metadata generation and processing audio data across GCP. Lead experimentation, model versioning, deployment, and user validation studies.
Software Engineer & Co-founder — GoodsLeads
Dec 2024 – Aug 2025📍 João Pessoa, Brazil
Developed a lead data catching and tracking platform using Google Maps API for targeted lead capture. Implemented LLM-based segmentation with Claude and LangChain for intelligent lead classification. Built automated email and content generation workflows using n8n. Architected scalable infrastructure on AWS for data processing and automation pipelines.
ML Research Fellow — Research Grant
Sep 2023 – Sep 2025📍 Salt Lake City, UT
Designed and implemented high-performing architectures for temporal audio synchronization with state-of-the-art with 5ms of accuracy and low latency. Built chords diagrams and chords extensions with Next/FastAPI. Published research findings in national Web Systems and MIR conferences.
Machine Learning Intern — Moises Inc.
Sep 2021 – Sep 2023📍 Salt Lake City, UT
Developed generative and classification models used by millions of users - semantic song labeling. Built scalable ML infrastructure with Docker and designed end-to-end data preprocessing pipelines for large-scale model training.
Fullstack Development Intern
May 2021 – Feb 2022📍 João Pessoa, Brazil
Created secure data portals and interactive dashboards for government transparency platforms using Angular and Flask REST APIs with real-time analytics.
Selected Publications
Representation Matters: Evaluating Spectral Representations for Temporal Modeling
📄 Cavalcante et al. • 2024
Proposed architecture improving temporal estimation accuracy by 26% on benchmark datasets.
HeroKeys: AI-Based Virtual Environment for Music Learning
📄 Jordão, Cavalcante et al. • 2024
Developed interactive platform combining ML-based performance analysis and gamified visualization.
Leadership & Contributions
Led integration of ML models into production systems impacting large user bases. Conducted product research and A/B testing for model improvement. Contributed to open-source projects in ML, signal processing, and AI education.