I'm an academic (researcher) and developer focusing on Julia Lang packages, Electron.js apps , social physics simulations 🏙️, and useful tools. Below, I've organized my key repositories by category. Check them out—many include detailed docs, examples, and ways to contribute.

These tested tools have been peer-reviewed by the Journal of Open Source Software (JOSS). Great for data processing, NLP benchmarks, and more.

• WunDeeDB.jl: A zero-config embedded vector database with WAL support. Useful for lightweight AI/ML embeddings. Ideal for vector search that scaled with transactions (which come with a cost for the confidence in data integrity) (JOSS-reviewed; in General registry)
  
  tip: using Pkg; Pkg.add("WunDeeDB")-try the examples for instant vector search.

• LMDiskANN.jl: Julia implementation of Low Memory Disk ANN for efficient nearest-neighbor searches. (JOSS-reviewed; in General registry)
  

  • tip*: using Pkg; Pkg.add("LMDiskANN") Ideal for large datasets and speed where transactions are not needed.

• KeemenaPreprocessing.jl: Preprocessing for text data: cleaning, normalization, vectorization, tokenization and more. (in General registry)

  

  tip: using Pkg; Pkg.add("KeemenaPreprocessing"): Great for many text preprocessing tasks and also allows it to perform with controlled memory when processing large text corpora.

• BenchmarkDataNLP.jl: Generates synthetic text (e.g., via Context-Free Grammars) with tunable complexity to test NLP methods like LLMs.

  

  Newcomer tip: Parameterize complexity to stress-test your models on less complex data before going into more complex natural corpora.

Two finished desktop apps for media tagging and organization. Built with Electron.js for cross-platform use—download releases for Linux/Windows.

• Tagasaurus: Tag your planet with semantic search and machine learning. Search and annotate your media files with ease keeping all the data locally.

• TagasaurusLegacy: The original "Tag your Planet" app—gateway to your semantic multiverse. Here memes and emotions take a first place position in search with specific features to allow for users to search for the impact of memes on certain keywords by utilizing a bi-partite graph.

Repos with code implementations from my research papers on social dynamics, entropy, and models like Schelling. Useful for simulations in physics/social sciences—many include Jupyter notebooks for reproducibility.

• Eurovision: Investigating the collusion between countries for the complete historical data trajectory of the Eurovision Song Contest
  featured in the economist magazine online and in print: https://www.economist.com/blogs/graphicdetail/2018/05/daily-chart-7 , https://www.economist.com/news/europe/21742117-continents-annual-singing-tournament-becoming-increasingly-partisan-neighbourly-voting . tip: Run the notebooks to visualize the one-way and 2-way (collusion) biases.

• EurovisionBias: Investigating the collusion between countries with both the positive and negative biases. So that the tendencies for countries to allocate significantly more or significantly less can be seen.
  tip: Run the notebooks to visualize the full range of biases.

• Schelling Consistent: Addresses the issue of a decreasing entropy trajectory in the canonical Schelling simulation by introducing a dual dynamic with monetary exchanges so the arrow of time is respected.

• Rat Utopia Modeling: Explores how the population collapse in the Rat Utopia experiment (Universe 25) can be simulated through a degeneration of the entropy of the hierarchy given the full visibility the rats had in the enclosure.

Handy packages and templates for everyday productivity, like LaTeX setups.

• Nighttime LaTeX Template: A sonar themed dark-mode friendly LaTeX template for documents—reduces eye strain at night. Newcomer tip: Copy the .tex file and compile with pdflatex; perfect for late-night writing.

• UCF Style LaTeX Slides Template: Beamer template in University of Central Florida style for presentations.
  Newcomer tip: Customize the department logo.

• Star ⭐ or fork repos you're interested in!
• Open issues/PRs if you spot improvements.

Thanks for visiting! 👏