This is program designed to automate trading public equitieswith no user input. The program uses a deep reinforcement learning model to predict when it should buy, sell or hold a stock. The proprietary RL model is trained on approx. 2 years of data from alphavantage. Crucially, using an RL model allows me to modulate the risk profile of the program with a few minor parameter changes. The RL Trader scrapes a finviz.com screener to find the best stocks to trade on the day and uses interactive brokers API to place orders.

The AgriPod is a product which maximises fertiliser efficacy for farmers by advising them on the best times to apply fertiliser. It was designed by a team of three with me in charge of software. We've designed the AgriPod for developing nations such as India where farmers don't have access to costly, infrastructure-intensiveequipment found in developed nations. Our research suggests that we could nearly doublefertiliser efficiency. To learn more, visit the AgriPod website.

My Extended-Project-Qualification is a research project entitled Can an automated hydroponic system replicate or exceed the quality of herbs grown in their native region?. To explore this topic, I've designed and produced a deep-water-culture system and programmed an esp32 to optimise key conditions such as temperature and nutrient conditions. As well as this, I've produced a server so that the esp32 can log data measured from the system to ensure the ideal conditions are maintained.