Ibaad Rehman

Defined operational states of wind turbines based on clustering of historical SCADA data followed by classification of turbines into those states. Tokenization of scanned service reports exploiting structure of the reports identifying and retrieving key data. Developed a service protocol for maintenance of turbines based on historical error data by deploying deep learning models

Worked part time as data scientist at Cologne based startup Deployed computer vision algorithms to extract digits from fast changing terminal displays in assembly lines in order to digitize production records. Demand forecasting from historical sales data using deep learning models to minimize waste of perishable items.

Lead an effort for a funding proposal for a project on development of API for software defined radio implementation of a multi-standard radio. Worked as project lead managing a team of 5 engineers and 4 researchers for four years. Lead the developed of an API for physical and link-layer protocols of 802.11n and 802.16 communication standards for software defined radio implementation. Successfully demonstrated the execution of the project and delivered the project.

Designed and developed a locally manufactured trainer board as a product to be used for training communication and signal processing technologies. This product is used in communication labs in various universities in Pakistan. Worked with electrical engineering department to establish a state of the art brain computer interface laboratory. Designed training manuals and laboratory exercises to be used with trainer boards.

Developed simulations to access performance of various space-time coding techniques Conducted lab sessions for computer networks and computer architecture courses for undergraduate students

Thesis: A Deep Learning Model For Fault Prognosis In Wind Turbines: Developed a model for forecasting errors in wind turbines. The model leveraged autoregressive deep learning model to forecast power generation over a multi-step horizon utilizing SCADA data and exogenous environmental data. The model also utilized empirical power curves for the turbines in order to determine which forecasts indicated impending faults along with level of certainty for the forecasts.

Major: Computer Engineering Minor: Mathematics