IoT data analytics platform as a proof of concept containing state-of-the-art predictive maintenance algorithms. Developed during the Wind Energy Hackathon 2018 within 2 days and based on server-less technology and Python.
Technology demonstration: Image recognition service implemented using Inception V3 model implemented in Keras and Flask and served on AWS.
Technology Demonstration: Conversational Chat and Voice Bot talking about Ginkgo Analytics. Implemented with Amazon Lex, Transcribe and Polly.
Airbus AI Challenge
Unsupervised anomaly detection of sensor data from helicopters, implemented using a single class support vector machine algorithm (scikit-learn) and deployed with dash frontend on AWS cloud service.
Technology demonstration: Emotion classification using the Google Vision API. The front end is built in Python and deployed on a virtual machine the Google Cloud.
Visualization of eCommerce data in a dashboard to show incoming orders, revenue und marketing performance. Implemented in ElasticSearch and Kibana, hosted on AWS.
Car damage detection
Showcase which applies a R-CNN (Regional Convolutional Nezral Network) to images of cars and marks damaged areas. Useful for car rental companies, car sales platforms and insurances.
Legal Tech - D&O Detection
Machine Learning Model which marks deliverables and obligations in contracts. Smart service integrated into Open3 platform by StatusGreen.
Sales Cockpit PoC
Identification of best matching cars from online customer behavior for a leading car manufacturer.