Bardh Prenkaj

• Optimised API response time by a factor of two relying on a distributed hash map deployed in GCP to efficiently retrieve objects • Automatised support request creation on a large-scale Italian eco-friendly public service • Developed bug-free functionalities to satisfy the stakeholder requirements on three unicorn projects within the company • Assisted senior managers in making decisions on development time-frames based on technical complexities and allocated human resources

• Managed a team of 6 researchers in cutting-edge research in trajectory prediction of the novel coronavirus infectious trend • Assisted a team of 20 in developing an intelligent system capable of predicting anomalies in behavioural patterns in elderly people • Implemented ergonomic solutions of data analytics and data visualisation to aid healthcare personnel • Exploited automatic and intelligent models to cope with self-isolation problems of older patients in retirement homes in Lazio, Italy

• Taught Web and Social Information Extraction, a master’s degree course averaging 60 students per semester, covering the following topics: architecture of information retrieval systems, web information retrieval, social network analysis • Organised coding challenges in live lab sessions to incorporate applied machine learning to the theory provided in the course • Led a five-month project to detect shifting trends in AI topics to build a model with 93% of prediction accuracy

• Designed a NodeJS backend on Firebase to deploy APIs that expose standard and technical information of EVs to the front-end • Strengthened the data storage on an Amazon instance by relying on a MongoDB database fully updatable with new information from the stakeholders • Assisted project managers to assess the implementation effort of technical software functionalities • Led a team of 3 junior software developers in building complex frameworks of APIs to provide a continuous service to the stakeholders

• Developed a novel boosting-based autoencoder ensemble system to detect data anomalies with an improvement of 6% over the current systems in the state-of-the-art • Collaborated with two senior researchers from George Mason University (US) and an associate professor from the University of L’Aquila • Scheduled weekly virtual presentations in accordance with the project goals incrementing the productivity by 13%

• Led the practical part of the course Social and Behavioural Networks, a master’s degree course in Data Science with 50+ students attending regularly • Addressed problems in crawling web pages loaded dynamically by relying on Selenium and PhantomJS • Engineered web crawlers capable of surpassing IP blacklisting mechanisms via layered routing on Tor • Designed a Lucene index to persistently store web pages to provide a caching mechanism for online and offline crawling