Muhammad Ali
B.Sc.
+49 221-8275-4316
muhammad.ali@th-koeln.de
Sprechstunden
Working hours
Montag, Dienstag, Mittwoch, Donnerstag, 10.00 bis 18.00 Uhr
Campus Deutz, Betzdorfer Str. 2, Altbau, Raum HW2-66
Please make an appointment via e-mail
Funktionen
- Researcher
- Data Scientist
- Data Engineer
- Cloud Computing
Aufgabenbereiche
- Cloud computing
- Machine Learning
Forschungsgebiete
-
Cloud Computing AWS, Azure I specialize in developing cloud-based pipelines for building automation systems, focusing on real-time data ingestion, time-series forecasting, and anomaly detection. By leveraging tools like Apache Kafka, AWS (MSK, S3, SageMaker), and IoT platforms, I e
-
Data Science: Data Scientist specifically working on optimisation of setpoints in real-world systems using Artificial Intellegence and Machine Learning.
Projekte / Kooperationen
-
AWS pipeline for buildings data analysis and machine learning
Description: Designed and implemented a cloud-native pipeline to support real-time data ingestion, storage, preprocessing, time-series forecasting, and anomaly detection for smart building systems. Key Responsibilities: Integrated Apache Kafka for scalable ingestion of sensor data (temperature, humidity, occupancy, etc.). Used AWS MSK, IoT Core, and Lambda for event-driven data handling. Stored raw and transformed data in Amazon S3 and Amazon Timestream for efficient querying and historical tracking. Preprocessed data using AWS Glue and orchestrated workflows using Step Functions. Developed forecasting models using SageMaker DeepAR and Prophet for predicting energy consumption and environmental variables. Deployed custom anomaly detection algorithms to identify abnormal patterns in HVAC and energy data. -
Artificial Intelligence in buildings setpoint optimization
Working on this project to control the buildings setpoints to minimise the energy consumption and to provide comfortable environment for individuals using artificial intelegence. Creating digital twins of the building. RL agent interaction with the twin -
Control building setpoints using AI
Using RL agent controlled buildings setpoints. Use modelica simulations for building. RL agent interact with the simulaiton. Then deployed this trained agent on a real system. -
RL agent vs PID
Implementation of RL agent on seesaw and also on water level controller to make it easier and more accurate, compared RL performance with PID control. Training the agent in simulation and then deploying it on a real system. Created ETL pipelines. Used databases to store the real time data from the system.
Lebenslauf
Data Science |