Pedro Novaes

Student Researcher @ Núcleo de Estudos em Sistemas Coloidais (NUESC)

About

I build AI systems from end to end — from raw data pipelines to models fine-tuned for production, served via APIs and monitored at scale. My background spans both backend engineering and machine learning, which means I don't just train models — I deploy them, integrate them into real systems, and keep them running. That full-stack perspective is rare, and it's what lets me ship AI features that actually work in the real world. What I work with:· ML/AI: Python, PyTorch, TensorFlow, HuggingFace (BERT fine-tuning), scikit-learn, NLP, Computer Vision · Backend: Java (Spring Boot, JPA, Security), FastAPI, REST APIs, Microservices · Infrastructure: Docker, GCP, RabbitMQ, PostgreSQL, MLOps pipelines · Frontend: React, Next.js, TypeScript 📩 pcarvalhonovaes26@gmail.com · 💻 github.com/Pedronovaes1

Country

City

Brazil

Industry

Computer Software

Skill

Data Engineering, Machine Learning, MLOps, Natural Language Processing (NLP), Computer Vision, Deep Learning, Python (Programming Language), Data Science, Reinforcement Learning, PyTorch, Docker, FastAPI, Architectural Design, Statistics, pinns, Engineering, Algoritmos de aprendizado de máquina, Aprendizado supervisionado, Aprendizagem profunda, Pandas (Software)

Experience

Student Researcher

Núcleo de Estudos em Sistemas Coloidais (NUESC)

2025-2 - Present · 1 yr 8 mos

Аракажу, SE

Research engineer at NUESC running three parallel projects in applied ML and software engineering — two of them in the oil & gas domain: • Applied ML research for oil & gas • CNN (PyTorch) for computer vision-based analysis of scaling/fouling in oil & gas pipelines, automating manual visual inspection. • Reagent & glassware management system (FastAPI + React) with full audit trail, replacing 5 manual spreadsheets — validated in a live demo with 30 lab users. • Performing data engineering and preprocessing on large experimental datasets with Pandas and NumPy. Stack: Python, PyTorch, Scikit-learn, Pandas, NumPy, FastAPI, React, TypeScript, SQLite, Git · Agile/Scrum

Full Stack Engineer

Freelance / Independent Projects

2025-7 - 2026-4 · 10 mos

Аракажу, SE

- Developed an interactive Gantt chart application using React.js to improve task tracking and project visualization for operational teams. - Engineered a Python-based API for DNA data management. - Automated business workflows using n8n, integrating databases and real-time notifications to reduce manual processing and improve operational efficiency. - Designed backend architectures and REST APIs to centralize data and enable scalable system integrations. Tools & Technologies: Python, Java. REST APIs, React.js, n8n, PostgreSQL, Workflow Automation, Git

Academic Industry Residency Program | Python | Machine Learning | Docker | RabbitMQ

Grupo Tiradentes

2025-2 - 2025-6 · 5 mos

Аракажу, SE

- Trained and fine-tuned a BERT-base-portuguese model, achieving 90% accuracy in sentiment classification for IT support tickets. - Developed a Python-based REST API with FastAPI to integrate the AI model, handling message classification and persisting data in a PostgreSQL database. - Implemented asynchronous messaging with RabbitMQ to ensure scalability and efficient parallel processing of sentiment analysis requests. - Created an interactive dashboard using Next.js and TypeScript to visualize sentiment metrics, generating strategic insights for the support team. - Utilized Docker for service containerization, simplifying deployment and environment replication across development stages. - Structured a modern, scalable, Cloud-Native architecture, ensuring system robustness and maintainability. - Managed version control with Git and organized the project using agile practices, promoting continuous delivery. Tools and technologies: - Backend: Python, FastAPI, SQLModel, API Rest - AI: Machine learning, BERT - RabbitMQ - Postgres, Docker, GCP - Git, Github

Frontend Developer Internship

Grupo Tiradentes

2024-8 - 2024-12 · 5 mos

Aracaju, Sergipe, Brasil

- Development of communication and teamwork skills, using agile methodologies. - Developed and managed Docker containers for efficient deployment on Google Cloud, reducing environment setup time by 15%. - Applied Scrum methodologies and Trello for task management, accelerating the project delivery cycle by 20%. - Implemented responsive design using Material UI, significantly improving accessibility on mobile devices. - Managed Git version control to ensure traceability, organization, and efficient team collaboration. Tolls and Technologies: - Backend: Typescript - Frontend: React - Infra: Docker - Agile Methodologies