B U I L D I N G A R O B O - A D V I S O R
Algorithmic Trading Strategies Using Machine Learning
For my final year project, I’m building something I’ve always been curious about: an AI-powered robo-adviser that helps people invest smarter by actually understanding who they are. Instead of relying on those rigid, one-size-fits-all portfolios you see in most digital wealth platforms, my project uses machine learning to create investment strategies that feel personal, adaptive, and genuinely aligned with each user’s goals. The idea is to make investing more transparent and far more tailored to the individual behind the numbers.
Everything starts with a survey that gets to know the investor. Rather than slotting users into generic categories, the system asks about their risk comfort, how long they want to invest for, their ethical and ESG priorities, and the sectors they’re most interested in. These answers are then translated into data the model can work with, shaping things like volatility targets, diversification levels, and screening rules. It’s a simple way of capturing a person’s financial personality in a format an algorithm can understand.
Once the system knows the investor, it uses machine-learning-based portfolio optimisation to build a strategy that fits. It can expand or shrink the investment universe, filter out companies that don’t meet ESG expectations, and adjust the weight of different assets based on what really matters to the user. Unlike static robo-advisers, this model learns from market patterns and continually aims to refine its recommendations, balancing return, risk, and personal preferences in a smarter way.
To make everything intuitive, I’m developing an interactive Streamlit dashboard that shows the portfolio in real time. Users can explore allocation charts, performance trends, risk and return metrics, and even see exactly how their responses shaped the outcome. My goal is to keep the technology powerful but also approachable, so people feel confident understanding the decisions the AI is making on their behalf.
This project means a lot to me because it brings together two things I’m passionate about: data science and human-centred design. It shows how machine learning can make investing more personalised and more inclusive without losing transparency. Looking ahead, I hope this work demonstrates how AI-driven tools can move beyond static wealth management and shape a future where investors have guidance that truly reflects their values, goals, and identities.