London, England, United Kingdom
Developing AI features that simplify processes for assessing market risk. Implementing different exchange margin models.
• Developed Penrose Inverse and FFNN models in TensorFlow to forecast end-of-day closing prices using various historical opening price windows extracted with yfinance. • Conducted research using Auto-Correlation Function (ACF) plots to identify and quantify underlying trends in stock price history, determining which stocks were suitable for linear regression and the extent of dependency on past values. • Demonstrated that the effectiveness of the historical window size varied based on the nature of the auto-correlation. For stocks with recent auto-correlation, a smaller window was more effective. In contrast, for stocks with a more extended historical trend in their auto-correlation, enlarging the historical window improved model performance. • Optimized models for accurate stock price direction prediction and presented my work.
• Tutored in Python scripting, Java Minecraft server modding, JavaScript Game Development as well as the Unity, Godot and Roblox game engines. • Oversaw the workload of three other tutors and technical set up each weekend or holiday camp. • Guided clients through software installations by video call.
• Developed a GUI application to visualise JSON market data in files using Tkinter. • Implemented a recursive algorithm to handle arbitrarily nested data and rip exchange rates into a tree view. • The project was incorporated into the in-house uqlx quantitative analytics library.
• Implemented a hypothetical AAA racing game prototype using Unreal Engine 4. • Experienced two cycles of professional agile software development; actively participated in scrum meetings, and engaged with mentors to scope a practical prototype. • Presented a demo and development plan to Senior Management