Uttam Dhakal

Worked on the WALDO (Widely Applicable Library for Detection of Outliers) project. The project was about finding a needle in a haystack. We created an open-source framework to detect outliers in large datasets. We focused on creating state-of-the-art machine learning techniques for anomaly detection, as well as custom-tailored data visualization. With modularity at its core, the project's architecture was developed with flexibility in mind.

Worked for a Trapeze project funded by European Union. Trapeze stands for TRAnsparency, Privacy, and security for European citiZEns. This project is developing technologies to empower citizens to actively contribute to the cyber resilience of the European data space. It provides the public with the necessary tools and information to manage their security and privacy. It also creates technologies to enforce the integrity and non-repudiation of citizens’ data usage policies and processing across data sources.

We built the app called TYDR (Track your daily routine) to collect all kinds of data from users (with their permission). We used machine learning on the collected data to predict the personality of the user. The user would be assigned one of the big five personality traits, i.e. openness, conscientiousness, extraversion, agreeableness, and neuroticism. We also analyzed their data to predict their home, work, and favorite location.

Worked in a self-sovereign identities (SSI) project funded by the ministry for economic affairs and energy as part of creating “Secure Digital Identities”. SSI technology offers flexibility and autonomy when identifying, authenticating, and authorizing persons, businesses, and objects. Thirty-five applications from everyday life were put into practice and tested in areas as diverse as eGovernment, finance, industry, IoT, eHealth, education, identity and access management (IAM), eCommerce, and mobility.

Created a dashboard to predict the network traffic for top autonomous systems such as Google, Amazon, etc. Network traffics were predicted with various statistical algorithms and deep learning models. The dashboard was created in Plotly Dash. The project also focused on data Visualisation and Data Analysis. In the Dashboard, users could choose how they want the predictions to be made. They could either choose the deep neural networks or the statistical model such as ARMA and SARIMA.

Created web-based learning units for distance learning courses offered by the Beuth University of Applied Science. The main tasks included creating the online content for the courses, creating quizzes and exams in the moodle, and organizing the course content.

Worked in a eDissPlus project (Electronic Dissertations Plus). It is a project from a German research group (DFG) focusing on electronic theses. In cooperation with the German National Library (DNB), conceptual and technical requirements were addressed in order to secure archiving and publication standards for dissertations accompanying research data.