David Fekete

TheVentury is an Austrian startup highly focusing on innovation, growth and execution. My main role was helping clients to develop the most effective data and machine learning solutions. From ML side I mostly worked with NLP models and chatbot like mechanism (custom chatbot, chatGPT). Also I actively took part in software architecture design, and application deployment. One of my biggest success during this time to be able to redesign a conceptually wrong application to actual working solution as its Lead developer.

Foundation is the subsidiary of Hungarian Bankholding. There was a huge fusion between three Hungarian bank. One of my main task to develop and re-design an ML system which is able to find the corresponding customers between this three bank. This functionality is necessary for legal reason, so it was a key task during the fusion. Finally I and my team was able to create a Siamese model which is able to deduplicate customer with 99,996% efficiency. Furthermore I had to hold different ML education classes to broad the knowledge about ML inside the bank. I also took part in solving different business problems, holding agile sessions, and project management.

Bosch is one of the greatest automotive industry all around the world. I worked at the Test-Bench-Development department. In the beginning of my work I had to create a Power-Supply-Unit controller in C. Later I take part in TBD in Labview, then I had to develop a motor noise analyzator neural network, which was able to tell is the motor wrong or not, and if so then where is the error. This was a super successful project, thus the AI domain was relocated to Budapest in TBD department. I also worked closely with manufacturing and I took part in project management and developing AI strategy. I was asked to develop an AI framework which is suitable to every test-bench. I designed it, then implemented it in Python, and I also prepared the system for multiprocessing on Hadoop with Spark.

Based on my experiences an IT company asked me to create a PoC to there new application in orthopedics. I had to produce an NN model which was able to localize different points on the patient on the video. There were multi-recordings, which was being recorded from different angles, and then based on these recordings it is able to map the points to a 3D coordinate system. The PoC was successful, so based on my results they invested to the development of a complete application.

Vidanet is subsidiary of Hungarian Telekom. During my work I had to take part mostly NLP tasks. In my first three month I created an NN model, which was able to detect, whether the customer is angry or not. To improve this process later I developed a Hungarian speech-to-text model and its complete pipeline. This model was an Autoencoder-LSTM model. It was able to create a transcript from the audio file, although it was needed also different post-processing to achieve reliable results. I also worked on image processing modeling, where the NN model had to localize digits on the router and digitalize it, to facilitate subcontractors work. This NN model was implanted to a mobile app.

Waberers is one of the biggest transportation company in EU. Beginning of my work there was a huge transformation in the firm. I had to create interactive workflow diagrams about processes, for education purpose. Later I had to take part in cleansing of employee's data, and data migration to Azure, and Active Directory. During this transformation I had to also take part in market research and project management. One of my final task was to develop an complete application with O365 tools, such as PowerApps, and MS Flow. This app made for the HR department to digitalize and automate recruitment processes, in accordance with national law.

• Maintained tight tolerances with design specifications.
• Assessed equipment after each production run and performed preventive maintenance to keep machines running smoothly.
• Moved finished work to correct staging area for disposition.
• Quality Control on produced workpieces