Works 📚
ML Skill Profiles: An Organizational Blueprint for Scaling Enterprise ML
Contributing Partner on the Applied AI Whitepaper, facilitating the delineation of essential skill profiles for successful implementation and scaling of machine learning initiatives.
Used:
- MLOps
- Organization
Content-based Music Recommendation (MSc. Thesis)
Recent boom in the industry of digital music distribution has made recommendation systems more of a necessity than a convenience both for consumers as well as business. Recommendations hitherto are mostly provided using collaborative filtering methods. This, however, fails to provide appropriate recommendations if no or sparse feedback data is obtainable which is known as the cold start problem. This research project aims to evaluate whether different machine learning techniques like CNN, RNN, TripletNetwork and transfer learning can overcome the cold start problem through content-based recommendations by predicting latent factors obtained by using a collaborative-based model. Quantitative and qualitative evaluation showed that machine learning based models are able to provide relevant recommendations and outperform traditional approaches and that musical-embeddings play a crucial role in doing so. Findings from this work can be used to provide better recommendations, perhaps leading to more business turnover and higher customer satisfaction.
Used:
- Python
- TensorFlow-GPU
- GCP
Blockchain Smart Contract Transfer Pricing Model (B.A. Thesis)
An overview of business management options using a blockchain smart contract pricing model to evaluate and set transfer prices while taking European tax regulations into account.
Used:
- SAP
- Solidity