Development of an Outlier Detection System for Usage Statistics
Background
Osapiens supports global companies from various industries in establishing sustainability
within their organizations and positioning themselves for the future. To achieve this, we
develop holistic Software-as-a-Service solutions that create transparency and sustainable
growth along the entire value chain, fulfill legal ESG requirements, and automate manual
processes.
We are leveraging our cloud-based multi-tenant platform, the ‘osapiens HUB,’ to help
companies seamlessly implement and automate compliance with international and
national ESG laws and guidelines, including CSRD, EUDR, and CSDDD, and to further ensure
accountable sustainability reporting. The osapiens HUB is designed to handle diverse
scenarios such as track & trace solutions, ESG compliance, and maintenance management.
With its robust infrastructure, the platform processes over 25,000 write operations per
second, enabling efficient data handling for a wide range of applications.
Project Description
The aim of this project is to develop an automated system that analyzes customer usage
statistics of the ‘osapiens HUB’ in real time and identifies outliers. Machine learning
methods should be employed to detect patterns in the data and pinpoint deviations. The
project must consider different usage scenarios, such as variations during the day, across
weekdays, and even seasonal patterns. Students will be tasked with developing a robust
and scalable solution capable of processing both current and historical data effectively.
Project Tasks
- Extract and structure usage data from the cloud infrastructure
- Preprocess usage statistics to prepare them for analysis
- Apply suitable machine learning algorithms for outlier detection
- Develop and implement methods to identify anomalies in various metrics
- Evaluate and optimize the model in terms of accuracy and performance.
- Implement an interface to integrate the system into the existing infrastructure.
Requirements
- Strong knowledge of data analysis and machine learning algorithms
- Experience with programming in Python and relevant libraries such as scikit-learn, TensorFlow, or PyTorch.
- Familiarity with large datasets and cloud technologies is a plus
- Independent working style, team spirit, and analytical thinking skills
We look forward to receiving applications from talented and dedicated students who are
eager to contribute to the development of an innovative outlier detection system for the
osapiens HUB.