AI Solution for technical cleanliness on medical materials

• I have carried through a data science project, that showed 30% of dental implants contaminated with particles, hazardous to human health and not compliant with requirements in clean production.
  
• I have created a AI solution for manufacturers, which facilitates automated image analysis. 
• Software development for the solution included product planning, python coding, integration of machine learning components to the backend, data base programming, development of micro services, testing and deployment.
• The data science project included project planning, data akqusition, image labelling, AI training, and deployment of the AI model.

• Machine learning model for identifying unreliable news articles from a labeled dataset
• Preprocessing steps include text cleaning, tokenization, lemmatization, stopwords removal and vectorization with TF-IDF
• Classification with numerous Machine Learning algorithms and evaluation

• Creation of a multi-agent system for autonomous agents, that interact in a shared environment to attain personal interests and/or to complete common goals
• Solving Distributed Constraint Optimization Problems (DCOP) in a framework that models the agents autonomous behaviour
• Coordination and communication according to the problems structure

• Build and train Text-to-lmage DCGAN framework trained on the Oxford Flowers-102 dataset
• Char CNN-RNN network is trained on visual and textual information and outputs an embedding vector that captures the underlying relationship of text blocks
• DCGAN model for generating realistic images based on text
• Results evaluated with Frechet Inception distance

• Identification of 1,251 celebrities, extracted from videos uploaded to YouTube
• Data exploration and cropping of each voice clip
• Extraction of MFCC, Delta MFCC, ZCR and LPC features
• Creation of a Neural Network and a Gaussian mixture model
• Performing data inference, evaluation and visualization

• Creation of A DCGAN framework that produces realistic 3D surface Datasets on the nanoscale
• The framework serves to create realistic data sets, which represent the nanoscale environments and the atomic roughness on that scale
• A combination of novel similarity contents are incorporated in the training procedure such as Height Histogram, Fast
• Fourier Transform and data discretization with N-gram graphs

• Designed and developed robotic systems that increased the production and precision level in industrial applications
• Developed and deployed software for pick and place Delta robots with Keba KeMotion robotics software and Codesys
• Investigated and conducted 2D and 3D machine vision for object detection with Cognex, Omron, Keyence and LMI image sensors
• Performed signal processing, motion planning, trajectory optimization, robot dynamics and requirements engineering analysis throughout the development stages
• Executed integration testing, evaluation, quality control and commissioning

• Designed and developed an industrial robotic system as the main software engineer for a crucial business change project
• Developed and deployed software for PLC, SCARA robots and HMI with Siemens TIA portal, Simotion Scout, WinCC and Epson RC+
• Played a key role in formulating the rotation and translation mathematical algorithms used for object manipulation
• Researched and accomplished machine vision with Cognex image sensor
• Set up communication protocols among hardware components, performed signal processing, motion control and product tracking
• Applied integration testing, evaluation and commissioning