My name is George and I am a data scientist. I always enjoyed automating processes, working with data, run simulations and doing science.
I am passionate about new technologies and methods, love to learn every day new things and apply them to my work. I enjoy doing research, and working on proof-of-concepts, as well as developing innovative softwares.
Starting as an engineer, with specialization in Robotics and Control Theory, I grew to become a data scientist, with strong background and experience in Deep Learning and Computer Vision.
However, I am always willing to expand my knowledge in Machine Learning and apply it in Computer Vision, as well as Natural Language Processing.
Currently, I am working as a Data Scientist in Connected Services at Fujitsu Germany.
Modules in Computer Science, Applied Mathematics and Scientific Computing
Application Areas: Data Driven Computing, Big Data, Machine Learning, Deep Learning, Computational Physics
GPA : 1.6, in German System
GPA : 8.50
Major: Mechanical Design, Control Systems & Robotics GPA: 9.05
Part of the Delivery team of Fujitsu Connected Services, focusing on solutions in the automotive and manufacturing industry. Main tasks include the use, maintainance and further development of Fujitsu's AutoML system, especially for supporting and automating the visual inspection pipeline. Moreover, custom deep learning and machine learning solutions are developed for extending the AutoML platform. Working on the sandbox development of Fujitsu's Digital Manufacturing Platform. This is a solution towards smart factory, i.e connecting different levels and parts to improve the production performance. Among other tasks, I automated the collection, parsing and communication of factory data from a simulator to a database through MQTT protocol for performing analytics and visualization.
Responsible for the virtual advertising project. Combination of classical computer vision methods, like image registration, developed in OpenCV, with deep learning methods, mainly Mask-RCNN, in Caffe2, to segment and produce highly accurate advertisement masks.
Master Thesis, Automating the generation of a synthetic dataset, by superimposing objects on real images with the use of a computer graphics software, Blender. Image registration, depth completion and 3D road segmentation were implemented in Python. The created dataset was used to train a deep learning model to perform lane detection and performance was compared with real datasets, written in Keras.
Responsible for setting up and maintaining company’s database, deployed in Docker. Migrating existing MySQL database to MongoDB. Responsible for collecting data from various sources and APIs and storing them. Daily, weekly and long term data analysis, regarding different relevant KPIs was performed with Python, Pandas and Tableau to increase productivity and suggest improvements.
Diploma Thesis, Applications with Robotics software and hardware. For this purpose ROS (Robot Operating System) was used, which is supported by Linux. Theoretical analysis and real-time implementation of game theory to the robotic platforms, Asctec Firefly and Pioneer Mobile Robot. Applied sophisticated controls, such as model predictive control, written in C++.
Interested in all the types of ML, supervised, unsupervised and reinforcement. Enjoy working with structured and unstructured and apply different learning methods for a variety of tasks.
Interested in applying end to end DL solutions to different problems. From data generation, exploration and cleaning to training, validation and deploying models. Really enthusiastic in researching new architectures and methods.
Deeply interested in CV and making computers able to see and understand. Very keen on traditional CV approaches, as well as newer DL approaches, for solving problems like object detection, instance segmentation etc.
Robotics have always fascinated me. Combining DL, CV, Control Theory and Automation, while using different sensors is a process I find exciting working on.
Enjoy applying simulation of systems coming from different sciences, such as physics, molecular dynamics etc. Interesting in developing and applying numerical algorithms, along with scientific computing, and numerics to solve complex problems fast and efficient.
Interested in all the challenges and goals of AI, thus increasing the capabilities of the machines and working towards the problem of general AI, by combining multiple engineering and science fields.
Click to read more
Click to read more
Click to read more
Here there is a list of my published papers so far.
G. Tzannetos, P. Marantos, and K. Kyriakopoulos, in IEEE, on the 24th Mediterranean Conference on Control and Automation, vol. 2016.June 21-24th 2016