Muhammad Shoaib Sikander

*Development of RESTful & gRPC based AI services and their integration in Python & C# applications * Exploration of Generative AI and creation of Python based RESTful service for Large Language Models * Writing docker-compose, Dockerfiles and Shell scripts for smooth deployment of Docker containers * Managing code in Git and implementing Bitbucket Pipelines for CI/CD processes * Collaboration with other teams through meetings and comprehensive documentation in Confluence

In this project from Deutsche Bahn (DB), the time series data from numerous sensors installed in an engine is collected and used for predictive maintenance. Firstly, the correct and malfunctioning data is classified. After that, statistical methods (ARIMA, SARIMA) and neural networks (LSTM, MLP) are used for univariate and multivariate time series forecasting and anomaly detection. Finally, a Meta-Learning Engine is prepared for automatic selection of the best Machine Learning model.

The goal of Master Thesis is to model, simulate and control a meshed structure District Heating Network such that every user, even on the farthest end of network, is able to extract its required heat from the network. Initially, the network is modelled in Simulink, but due to its limitations in handling meshed networks, a Simscape model is developed. Finally, MPC and PID controllers are applied to control the whole network. Similar concept can be extended to electrical circuits as well.

This work includes Image Processing for object detection and tracking using MATLAB. In this project, some droplets of an organic liquid are ejected into a chamber where they are exposed to flame, due to which these particles change their moving speed, diameter & area and eventually evaporate after a little time. Tracking the positions of these droplets and finding the change in their speed, diameter and area are the main tasks for this project.

The Python implementation of machine learning models takes more execution time, which is reduced in this project (for RaspberryPi) by implementing in C++. First of all, performance (accuracy and execution time) of different ML (DT, RF, SVM, KNN) and DL (Neural Networks) models is compared by implementing in Python. After that, two best performing models (SVM, KNN) are selected and their performance is improved in terms of execution time by implementing them in C++ using the concepts of OOP.

In this project, two pairs of stereo cameras with wide baselines (horizontal distance between two cameras) have been calibrated using multiple calibration objects (e.g. Chessboard, ArUco Markers, ArUco Board, ChArUco Board and Red coloured circles) and their results are compared. Calibration results have also been compared by moving or by fixing calibration objects in common field of view of both cameras. Finally, calibration results have been used to reconstruct images using Image Rectification technique.

To perform different tests on a metal sheet, a platform is moved under a fixed laser beam. TwinCAT 3.0 is used to program Beckhoff PLC to automate the platform in three different patterns i.e. Manual movement, Automatic zig-zag pattern movement & Step movement with possibility of steps of different distances. Programming is done mainly in Structure Text (ST) and Function Block Diagram (FBD) for 2D movement of platform. Designing GUI to control the movement is also one of the vital tasks.

In this full time job, I worked in the Electrical Design Department of a construction company to develop Coordinated Shop Drawings from Issued For Construction (IFC) Drawings. These drawings must be designed keeping in mind the actual working conditions and taking care of other services installed on the site i.e. Fire fighting System, HVAC System, Normal Lighting System, Emergency Lighting System.

In this project, an automation and control system was designed using Arduino UNO Controller and GSM module. Different sensors installed in this system send their signals to the Arduino, which interacts with the GSM module Sim900d and sends text messages on operator’s mobile phone. The operator can send commands from mobile phone using text messages to turn on/off LEDs, to switch on/off relays and to control a stepper motor.

Process Automation-I & II Robotics-I & II Sensors & Measurement Systems Microsystems Control Theory-I Real Time Systems-I Advanced Digital Systems Design Process Automation Laboratory Control Laboratory