Ing. Khurram Shehzad

Entwicklung praxisnaher KI- und Softwaresysteme: In DIfA 2.0 entwickle ich eine multimodale RAG-Pipeline mit LangChain, FAISS, Qwen2-VL und NetworkX-Wissensgraph. Für DeTail realisiere ich eine YOLO-basierte Computer-Vision-Lösung sowie ein Full-Stack-Dashboard mit FastAPI, GraphQL, PostgreSQL, React und Docker. In NORDLINK integrierte ich HPC-, S3- und VM-Ressourcen, automatisierte SLURM-Workflows und implementierte Monitoring mit Prometheus und Grafana.

Projekt: ADSPEED - Adaptive Low-Latency Speech Decoding Training generativer Deep-Learning-Modelle (GANs, VAEs) auf Sprachdatensätzen zur Synthese von Sprache aus dekodierter Hirnaktivität. Transkription erzeugter Sprachsignale mit ASR und strukturierte Datenextraktion mittels NLP-Methoden.

Topic: Using Reinforcement learning to compute State dependent PI State Space Controller for Non-Linear System. Skills: Reinforcement Learning · Control Theory · Stable baseline · SAC · PPO · Machine Learning · Robotic perception · Artificial Intelligence (AI) · Image Processing · Python (Programming Language) · TensorFlow · Python · PyTorch · MATLAB · Linux

Project: GABRIEL ( GAnzheitlicher, HyBrid-ElektRIscher AntriebskomponEnten für die Luftfahrt ). Ausrichtung von 3D-Punktwolken mittels ICP- und H-Matrix-basierter Registrierungsverfahren zur Bauteilprüfung. Entwicklung Python-basierter Tools zur Integration von MATLAB- und C++-Skripten sowie Aufbereitung großer Ergebnisdatensätze mit Pandas.

Project: SeMoSys (Virtuelle Sektionsmontage und Systeminstallation). Aufbau einer 3D-Rekonstruktionspipeline aus 2D-Bilddaten und Ensenso-Stereokameras inklusive Kamerakalibrierung. Entwicklung YOLOv5-basierter Objekterkennungs- und Anomaliedetektions-Pipelines; Deployment mit Docker, Workflows in GitLab CI/CD.

Topic: Computing multiple input and multiple output State Space Controller using SAC implementation of Reinforcement Learning. Skills: Reinforcement Learning · Control Theory · Stable baseline · SAC · Machine Learning · TensorFlow · Python · PyTorch · MATLAB

Topic: Robotics perception of Scan robot for retail stores. Weiterentwicklung ROS-basierter Computer-Vision-Module in C++ und Python für robotische Wahrnehmung auf Linux-Systemen. Entwicklung und Training von Mask-R-CNN-Segmentierungsmodellen (TensorFlow, PyTorch) inklusive pixelgenauer Datenannotation.

"Real Time instance segmentation using yolact++ in pytorch" • Research Work and Documentation on instance segmentation. • Implementation of Yolact++ algorithm using pytorch for instance segmentation on annotated data set. Skills: 3D · Image Processing · Deep Learning · C (Programming Language) · Git

"Green AI: Efficiency Technology for Vegetative Plant Production" • Annotation of images with semantic segmentation model. • Annotation of images using f-BRS interactive segmentation model. Skills: Image Processing · Deep Learning · C (Programming Language) · Git

•Pulsed Excitation of Thermal Flow Sensors for Reduced Power Consumption and expanded measurement Range. •Research Work and Documentation. •Developing program for simulation in LABView. •Designing of 3D airflow medium using Autodesk Inventor. •Developing a new calculation method for compensation of air flow using Python NumPy. Skills: Image Processing · C (Programming Language)

"Smart Solar System using IoT Technique ”. •Developed and implemented a solar cell system in university that work autonomously and implementation of IoT technique to control the parameters of system using android application. •Development of android application. •Data monitoring of system also done using IoT technique i.e. updated data were transmitted to central server local server for storage and analysis then further more transmitted to mobile application wirelessly. Tools: Arduino mega, C language

• Siemens and Allen Bradley PLC with RS logix software • AC and DC Drives. • AC Motors, DC Motors, Three Phase AC Motors, Single Phase AC Motors, and Servo Motors • Different Sensors and Actuators. • Distribution of electricity through generators to different department

It includes the Subjects of robotics, machine vision, automation, control engineering, microsensors and Embedded systems. Apart from these heavy duty courses I had undergone through area of speech and signal processing, Presenting Science and Intellectual Property, Patents and Further Protective Rights.

The Diversity of this course has enabled me with the ability take account both aspects of the proposed system; software-level and Hardware-Level. The courses were Robotics, Computer Vision Modelling and Simulation Power Electronics, Object oriented Programming (OOP), Turbo C, C++, Embedded Systems, Digital Logic Design (DLD), Control System, Computer Networks, CAD, Internet of Things. Specialization: Robotics, Embedded System, Control Engineering and IoT.