Implementing an RCNN to detect traffic signs in images. In a previous project I have used an off-the-shelf library to detect faces in images, this time I wanted to build the dectection engine on my own.
As Cryptocurrencies and Blockchains are a hot topic at the moment, I decided to take a look into how dApps on the Ethereum Blockchain work and how to implement an ERC20 token (the standard for cryptotokens such as DAI, USDC or COMP). In addition to the token, I wrote a Smart Contract that allows interacting with the token based on fixed rules, similar to how DeFi apps such as UNISWAP or COMPOUND operate.
System that received images from a security feed (= webcam). It extracts and recognizes faces in the images and displays the data to the user revealing known and unknown faces.
Webapplication to store links to read in a common place. Working with Typescript, Python/Flask and Mongo DB. Using Facebook as service to authenticate users.
Rewrote the backend using Golang. Goal was to learn the language and improve the performance of the backend: new backend is faster (latency: ~86ms), lower storage footprint (Python Docker image: 960MB, Golang Docker image with a staged build: 19MB).
Develop an AI to steer an agent through the Flappy Bird game. Using an Evolutionary Algorithm to find the best set of weights of a Feedforward-Neural Network to improve the performance of the agents over time.
Webapplication to manage and view sensors. Sensor values are published to a Mongo DB backed service, the frontend dispays the sensor data. Access control using JWT.
Implementation of basic functionality of an operating system: task scheduling, memory allocation, handling keyboard inputs (interupts), bootloader to start the OS.
Building a Neural Network based classifier that predicts the outcome of a match in the Bundesliga.