Matthias Denecke
Verified Expert in Engineering
Generative Pre-trained Transformers (GPT) Developer
Matthias is an analytical, highly adaptable computer scientist with extensive experience in machine learning, artificial intelligence, and natural language processing. He is skilled in providing guidance to resolve complex technical challenges. As a technical lead, he has a track record of success architecting, developing, and deploying commercial solutions using artificial intelligence.
Portfolio
Experience
Availability
Preferred Environment
Speech Recognition, Scikit-learn, Keras, TensorFlow, Python, C++, Linux, Windows
The most amazing...
...project I've developed is a technical manual application guiding field service technicians through complex procedures by voice control only.
Work Experience
Freelance
Nano, Ltd.
- Integrated existing 2D and 3D video processing packages (Detectron, VideoPose3D) into a web app.
- Created a module detecting the potential for back injuries based on medical models, which is capable of processing the information extracted by the video processing software.
- Created a module capable of writing a natural language report detailing the potential risk for back injuries.
Consultant (freelance)
Nano, Ltd.
- Evaluated various deep learning models for natural language understanding.
- Developed deep learning approaches for named entity recognition, text summarization, and document retrieval.
- Integrated trained deep learning models in Flask-based web applications.
Lead Developer (freelance)
Kextil, LLC.
- Architected and implemented an app running on Windows, iOS, and Android, allowing field technicians in hands-busy, eyes-busy situations to interact with complex procedures and workflows through voice only.
- Conducted user tests in the field and improved voice interaction based on user feedback.
- Showed reduced task completion times by almost 20% when using the app.
- Maintained a cross-platform code base by separating application logic from UI and implemented UI using platform-specific APIs (Java, C++, Objective-C).
- Used standard collaborative tools such as Bitbucket, Git, and Jira for progress tracking and project management.
Consultant (freelance)
Ford Motor Company
- Designed a voice-operated system for in-car control of infotainment systems and seats.
- Developed speech recognition and natural language understanding resources.
- Evaluated systems through user studies and improved performance.
Research Associate
NTT Communication Science Laboratories
- Implemented a Restricted Domain Question Answering System based on NTT's proprietary QA technology in Japanese.
Research Associate
NTT Communication Science Laboratories
- Developed Japanese bus information system.
- Developed reinforcement learning algorithms for spoken dialogue management.
- Evaluated system with Japanese users.
Experience
Voice-Operated Interactive Technical Manual
https://www.youtube.com/watch?v=dWeeqxrbDLE&t=55sSkills
Languages
Python, C++, C, C#, SQL, Java
Libraries/APIs
TensorFlow, Keras, Matplotlib, Google Speech API, Scikit-learn, SpaCy, PyTorch
Tools
GCC, Visual Studio
Other
Deep Learning, Speech Recognition, Natural Language Understanding (NLU), Natural Language Processing (NLP), Artificial Intelligence (AI), Artificial Neural Networks (ANN), Text to Speech (TTS), Machine Learning, Unsupervised Learning, Supervised Learning, Dialog Systems, GPT, Generative Pre-trained Transformers (GPT), Reinforcement Learning, Recurrent Neural Networks (RNN), Explainable Artificial Intelligence (XAI)
Frameworks
Flask, Android SDK
Paradigms
REST, Cross-platform, Management
Platforms
iOS, Android, Windows, Linux
Education
Ph.D. in Computer Science
University of Karlsruhe - Karlsruhe, Germany
Master's Degree in Computational Linguistics
University of Stuttgart - Stuttgart, Germany
Master's Degree in Computer Science
University of Karlsruhe - Karlsruhe, Germany