Skip to content
View luizmlima's full-sized avatar
🎯
Focusing
🎯
Focusing
21 followers · 27 following

Highlights

  • Pro

Block or report luizmlima

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
luizmlima/README.md

👋 Hello, I'm Luiz Lima

💻 Computer Science student at UFCG, fluent in English, with experience in software development, Artificial Intelligence, automation, and Open Source development. Recognized in national scientific olympiads, with interests in cybersecurity and emerging technologies.

🔎 I am an Apache Lucene contributor, with a performance optimization accepted and merged into the official project. I have experience with Java, Python, C#, Machine Learning, Linux, Git, and software performance analysis.

I am especially interested in Software Engineering, Artificial Intelligence, Cybersecurity, Algorithms, Data Structures, and emerging technologies, and I am currently seeking opportunities where I can apply my knowledge and continue growing as a software developer.

🌐 Open Source Contributions

🔎 Apache Lucene Contributor

I contributed to Apache Lucene, a high-performance Java search and indexing library used at the core of technologies such as Elasticsearch, OpenSearch, and Apache Solr.

My contribution focused on the JapaneseCompletionFilter in the Kuromoji module, replacing an ArrayList used as a FIFO queue with an ArrayDeque. This changed repeated removals from the beginning from O(n) per operation to O(1) amortized.

I also built a JMH microbenchmark to compare both implementations and validated the change using Lucene's test suite.

The contribution was reviewed, approved, and merged into the official Apache Lucene codebase.

🔗 Apache Lucene PR #16416

📂 Open Source Optimization Project

🚀 MEISSA Project (LIAD/HP)

I participated in the MEISSA Project, an Artificial Intelligence training initiative resulting from the partnership between the Laboratório de Inteligência Artificial e Arquiteturas Dedicadas (LIAD) of UFCG and HP Inc.

The program focused on advanced topics in Machine Learning and Deep Learning, including:

  • ML Fundamentals: Probability, Statistics, Python for Data, and MLP.
  • Deep Learning: Deep neural networks and Transfer Learning.
  • Time Series: Statistical modeling and Deep Learning (N-BEATS, PatchTST).
  • Foundation Models: Embeddings and state-of-the-art models.

🏷️ Technologies & Tools

Java Python C# JavaScript PyTorch Scikit-learn Hugging Face Linux Git GitHub

📂 Featured Projects

🔧 Open Source & Software Engineering

Project Description Repository
Apache Lucene Contribution Optimization of FIFO output buffering in JapaneseCompletionFilter, replacing ArrayList with ArrayDeque. Accepted and merged into Apache Lucene. PR #16416
Open Source Project Optimization Analysis of data structure optimization opportunities in Apache Lucene, JavaParser, I-Educar, and Java Photo Gallery, including benchmarks and Pull Requests. Repository

🌟 MEISSA & AI Projects

Project Description Repository Link
Multivariate Forecasting MEISSA Final Project using PatchTST and AutoML for macroeconomic data. Forecast Colab
PatchTST Benchmark Systematic evaluation of Transformers on multiple temporal datasets. Benchmark Colab
Transportation Demand Comparison between N-BEATS and statistical models for vehicle trip volume. Transport Colab
Image Colorization with AI Transfer Learning (ResNet-18) to automatically colorize B&W images. Colorizer Colab
Fraud Detection MLP Neural Network with data balancing for financial transactions. Fraud Colab

💻 Software & Games

📫 Contact

Pinned Loading

  1. fraud-detector fraud-detector Public

    An AI-powered credit card fraud detection system using Multi-Layer Perceptron (MLP) neural networks and data balancing techniques to maximize recall and identify fraudulent transactions in imbalanc…

  2. image-colorizer image-colorizer Public

    An automated black-and-white image colorization model leveraging Transfer Learning with ResNet-18 and the Lab color space to reconstruct semantically coherent colors in grayscale images.

  3. multivariate-forecasting multivariate-forecasting Public

    Advanced multivariate macroeconomic forecasting using AutoPatchTST and AutoML on the FRED-MD dataset, comparing feature selection strategies and hyperparameter optimization.

  4. Optimization-of-open-source-projects Optimization-of-open-source-projects Public

    Aplicação de Estruturas de Dados para otimização de desempenho e memória em projetos open source (i-Educar, JavaParser, Java Photo Gallery e Apache Lucene). Inclui análise de complexidade, benchmar…

    Java 1

  5. time-series-benchmark time-series-benchmark Public

    A comprehensive performance benchmark of the state-of-the-art PatchTST (Patch Time Series Transformer) model across multiple diverse datasets and various forecasting horizons

  6. vehicle-trips-forecast vehicle-trips-forecast Public

    A comparative study of time series forecasting models, including N-BEATS Deep Learning and statistical methods (Theta, ARIMA), to predict transportation demand and vehicle trip volumes.