Veselin S.

SCADA is an application that visualises and records the status of automated hardware used in Ocado’s warehouses. It’s an event-driven streaming application, which processes events from real-time embedded controllers (PLCs, robots, robotic arms, video capturing devices and bots - which run on the storage grids) via a web-based UI. It supports engineers by helping to minimise downtime and improve the efficiency of warehouse operations.

Research project to evaluate the possibility of creating automated trading system using latest achievements in deep learning technology. The idea was to filter the market data and transform ticks to bars with equal body length - abs(Open - Close). If we can predict direction of the next bar with good probability, we can create automated trading system for financial markets. Several hundred GB of raw tick data data was used to prepare range bars (since 2000). Train, test and validation sets have been created. LSTM network was ideal candidate because it is proven to be good for time series predictions. The system was tested with major currency pairs, stocks, major indexes and crypto-currency pairs (only few years of data available for crypto). Multiple parameter sets and self-optimising techniques were used. The results were not very promising – we were able to achieve about 62% successfully predicted bars only for limited time frame usually when the market is quiet and for range bars that are quite small in pips. Calculating spreads and commissions into the equation this was not sufficient to create profitable automated trading system.