Good to hear you're enjoying the Machine Learning class!
Following a whirlwind term of labs, cases, reports and presentations over 10 modules at the University of Cambridge, I'd like to spotlight my favourite machine learning class by Dr. Alejandro Reynoso, where we explored financial applications of agentic ecosystems that can self-reflect, improve and execute algorithms iteratively. Here, we studied the basic architecture of an AI-driven financial workflow including LLMs, reasoning agents, action agents and data processors and learned to develop a multi-agentic, multimodal solution to generate text and translated audio reports based on factors like SMA, RSI, MACD, crossover signals, recent news and analyst reports from various sources. More broadly, we discussed: 💡The rising sophistication of analytical systems: from parsimonious econometric models approximating reality with a handful of parameters to transistors achieving remarkable precision by running billions 💡Emergence: the unexpected rise of advanced ML model capabilities such as reasoning due to model size, architecture and training, and parallels to biological leaps in evolution 💡Artificial general intelligence: the potential for machines to achieve human-level intelligence, particularly considering agentic ecosystems and genetic algorithms, which mimic evolution through self-selection, mutation, preservation and creation to solve complex problems I truly believe in the power of AI to make finance smarter, faster, and more inclusive through hyper-personalization, predictive analytics, and automated decision-making in intelligent ecosystems that adapt and evolve, much like the markets they aim to master. Going forward, I'm keen to, more deeply, explore ways to harness this potential while navigating important challenges of trust, ethics, and regulation. Cambridge Judge Business School
