Simple Blockchain July 2025
Lightweight, per-user virtual machines that start as you SSH in.
A blockchain built in Python with a focus on clarity and fundamentals. Each block is linked with SHA-256 hashes and mined using a proof of work system where valid hashes must begin with a fixed number of leading zeros. The system supports adding transactions, block validation, and a full chain integrity check.
I wrote utilities to display the chain in a readable way, verify consistency, and dynamically adjust mining difficulty. The project demonstrates the core principles behind Bitcoin and Ethereum immutability, consensus, and tamper detection, without the overhead of a full cryptocurrency stack.
This was a personal project to discover blockchain mechanics, but it doubles as a lightweight educational demo of distributed ledger concepts.
Links: GitHub,
Porsche Chatbot June 2024
A simple NLP and Neural Network personal project that sheds light to the beauty of German Automotive technology.
A conversational chatbot powered by natural language processing and a neural network trained on custom Porsche related intents. The system uses tokenization, lemmatization, bag of words, and a Keras sequential model to classify user queries into categories such as greetings, model history, fastest cars, and pricing. Based on predicted intent, the bot selects responses from a curated knowledge base about Porsche’s history, models, and record breaking vehicles.
The project demonstrates the fundamentals of intent classification, neural network training, and real time inference in a CLI chatbot environment. While it’s far from the most accurate chatbot, its predictions can sometimes be fuzzy, I love it because it reflects the joy of building something from scratch and seeing a working system come to life. Beyond being a fun demo of Porsche trivia, it also serves as a practical prototype for domain, specific Q&A systems built on lightweight machine learning.
Link: GitHub
