Tassos Mousoulides

• Started development on an online metaverse project using Next.js, Babylon.js and Express.js (TypeScript) with MongoDB.
• Utilized Blender for 3D modeling and animations, including writing Python scripts for task automation.
• Leveraged ChatGPT for code generation and problem solving, accelerating progress across newly entered domains.

During this year, I aimed to improve and better round my web development and game development knowledge, gaining more experience in JavaScript and Node.js and learning to utilize React.js and Next.js as front-end technologies and Express.js as a back-end technology, while also learning the Babylon.js and Three.js web interactivity engines.
Development started for a fairly ambitious online metaverse-related personal project, utilizing Next.js as the front-end with Babylon.js, a game engine meant for the creation of complex 3D and VR experiences right on the browser utilizing WebGL or WebGPU, and Express.js with MongoDB for the back-end, all written in TypeScript.
Due to the necessities of the project, I also learned and extensively utilized the Blender 3D modelling and animation suite to customize models, textures and create more than a hundred handcrafted and key-framed transitionable animations for characters. I also sped up a lot of the arduous processes by writing task-automation Python scripts for Blender.
Due to getting into many new domains, I also learned to efficiently use ChatGPT for code segment generation and problem solving, which was cross-validated with my own solutions, speeding up progress in many of these domains significantly.

• Automated the trading dataset's creation using Python scripts with APIs.
• Clustered and analyzed data to derive insights.
• Researched and implemented AI systems to guide traders, using supervised, unsupervised, and reinforcement learning models.
• Designed the company's dashboard using Plotly's full-stack Dash framework, providing user-friendly data visualization.
• Contributed to website development using HTML, CSS, and Bootstrap for investor presentations.

During my time at TradeGenie, starting with its conception, I helped flesh out ideas and test how feasible they were given the current leaps and innovations in deep learning research. The company's vision was to create software that would help traders avoid extreme losses and provide guidance based on behavioural and other characteristics, as derived from the data.
As an Artificial Intelligence Researcher & Engineer, my primary focus was to collect trading data, analyze it and build a basic version of our AI systems for demonstration purposes. To that end, I wrote Python scripts to automate the generation of such a trading dataset, using the API of an algorithmic trading platform, utilizing robo-traders, to collect, cluster and organize the data and to find behavioral similarities between those clusters, given engineered evaluation metrics. Unsupervised learning was used in the dataset while supervised and reinforcement learning models were designed.
As a Web Developer, I helped design and build our company's dashboard, which would serve as an intuitive point of interaction between the system and the user, giving information in an organized and easy to digest manner about behavioral characteristics, guidance on achieving targets, dynamic warnings and performance analysis while gamifying the experience. To that end, Plotly's Python library DASH was utilized, an integrated full-stack framework ideal for creating dashboards with interactive plots and diagrams. In addition, I programmed the first versions of the company's website, using HTML, CSS and Bootstrap, which were shown to investors.

• Responsible for assisting students who were doing the coursework for the undergraduate “Introduction to Databases” module, resolving questions and finding or helping in finding bugs in SQL code/queries during the lab sessions.

• Archived many of the laws and court cases of Cyprus in a database accessible to subscribing lawyers.
• Typed, corrected and validated the papers that came from the courts using OCR before archiving.
• Added manual keywords and information about each paper into the system in order for searches by clients to be made easier.

• Helped the programmers of Quorum, the business solution that the company had developed, with various tasks.
• Provided support via phone to clients of the software.