Renan Florencio

Machine Learning & AI

On my way to become an expert in machine learning and artificial intelligence.

Campinas, SP

Email LinkedIn GitHub

About

Machine Learning & AI researcher

B.S. in Computer Engineering at Universidade Estadual de Campinas (Unicamp), expected July 2027.

Research at Unicamp on synthesis of inertial data for driver behavior classification (CNPq / H.IAAC), and at Purdue University on LLM-based agents for optimization and formal reasoning (PONTES Mobility Program).

Interests: ML, LLMs, applied research.

Education

B.S. in Computer Engineering

Universidade Estadual de Campinas (Unicamp)

Campinas, SP · Feb 2022 – July 2027

GPA: 8.9 / 10 (Top 5)

Experience

Purdue University

Jan 2026 – July 2026

AI for Optimization Visiting Scholar

West Lafayette, IN

•Developed an agentic pipeline to solve optimization proof problems using LangGraph and LangChain.
•Applied Prompt Engineering, Orchestration, Verification and Validation of agent responses.
•Wrote proofs in code using Lean4 programming language
•Solved a new problem in optimization by applying the agentic pipeline
•Received a Distintion award for the "Agents for Optimization Proofs of Equivalence" presentation at the Undergrad Spring Conference

Hub of Artificial Intelligence and Cognitive Architectures (H.IAAC)

Aug 2024 – Dec 2025

Machine Learning & AI Undergraduate Student Researcher

Campinas, SP

•Researching synthesis of inertial data for classifying drivers with multiple behavior patterns (CNPq, H.IAAC).
•Designed and evaluated inertial sensor data synthesis techniques.
•Developed Python scripts to automatically generate driver routines and behaviors using LLMs.
•Used Random Forest, SVM, and XGBoost to classify behavior, and t-SNE to visualize data.
•Published a paper "Fooling the Model, Failing the Road: Benchmarking Inertial Sensor Fidelity in Driving Simulators" at BRACIS/ENIAC (2025)

Projects

Simulated Driving Behavior

•Researching synthesis of inertial data for classifying drivers with multiple behavior patterns (CNPq, H.IAAC).
•Designed and evaluated inertial sensor data synthesis techniques.
•Developed Python scripts to automatically generate driver routines and behaviors using LLMs.

Pandasscikit-learnLLMst-SNE

Language Learning Agent

•Used Langgraph to create an Agent to help search for content in a required language.
•Designed the data schema and the information storage and retrieval.
•Implemented a multi-agent framework with an orchestrator.
•Created an UI to interact with the user

LangChainLangGraphLangSmith

Credit Card Fraud Detection

•Preprocessed data and applied machine learning methods to classify an unbalanced dataset.
•Handled class imbalance and evaluated model performance for fraud detection.

scikit-learnPandas

House Pricing Prediction

•Predicting the price of a house based on its features.
•Used Models, SVR and Ensembles.
•Evaluated the best model using metrics like RMSE, MAE, and R².
•Analyzed feature importance and compared the models.

PandasScikit-learn

Computer Vision

•Stitching a panorama using feature detection, matching, and image blending.
•3D reconstruction of a scene using several images.
•Live warning system for blind people using object detection.
•Analysis of generative AI models, LoRA and Stable Diffusion.

OpenCVLoRAStable Diffusion

Oportuni Application

•Application to help students find internships and jobs.
•Implemented a web application using Django to help students find internships and jobs.
•Used MongoDB, HTML, CSS, and JavaScript to create the frontend and backend of the application.

MongoDBHTMLCSSJavaScript

Titanic Survival Prediction

•Creating a machine learning model to predict whether or not a passanger on Titanic survived,
•The project goes from cleaning the data to training different models and evaluating the best one.
•Used Random Forest, Logistic Regression, and Support Vector Machines.

PandasScikit-learn

Numerical Calculus Methods

•Standard methods implemented by me to get zero of functions, least squares and interpolation.

PandasScikit-learn

Data Structure Problems

•Solved problems in C involving pointers, heaps, queues, graphs, linked lists, matrices, and more.
•Focused on correctness and clean implementation of classic data structures.