Amélie Medem

● Designed and built new data pipelines in Azure Databricks for ingesting data to/from product applications, Azure Datalake, and PostgreSQL databases. ● Ingest Datadog metrics into Databricks, integrate them with other datasets, and visualize the insights in Power BI. ● Built and optimized models to organize and structure data from diverse applications and sources for data users to consume. ● Develop and maintain Power BI and Databricks dashboards to visualize data insights, monitor pipeline performance, and ensure data quality. ● Enhanced codebase quality by implementing best practices and establishing robust CI/CD pipelines leveraging Databricks Bundle Assets, GitLab, and SonarQube. ● Develop and implement Data Contracts as a framework to monitor data models and define clear data specifications. ● Partner with business teams to identify their needs and deliver tailored data solutions that drive value.

● Design the data pipeline architectures for 2 new products: 1/ a data visualization application of SoLocal businesses about various information such as their paying or free services, the audience, the statistics on clicks or views, the reviews... ; 2/ a geographic search platform of SoLocal businesses. ● Build and maintain ETL processes for data flowing between databases, Elasticsearch, the data lake, and those applications. ● Set up an Elasticsearch search engine and build and maintain pipelines from data lakes to feed Elasticsearch. ● Improve and optimize the former data pipelines and SQL Stored Procedures. ● Use time series and deep learning techniques to analyze user clicks to predict the future and propose new services to businesses. ● Create thorough unit and integration tests. ● Train 10+ Amadeus employees on Scala.

● Develop a knowledge management platform Stample2 using technologies like Play Scala, React, Redux Typescript, and Elasticsearch. ● Develop from scratch a Web application in Scala dedicated to the management of psychologist patients for the Paris area (Qualisocial3). ● Ensures applications' ability to interact with multiple APIs and databases. ● Refactor 2000+ lines of code to migrate an older version of the application from BackboneJS to ReactJS. ● Do peer code reviews. ● Lead all the Research Tax Credit (CIR: Crédit Impôt Recherche) projects. This allowed Stample to get public research funds from 2014 to 2016 and to earn the JEI status (JEI: Jeune Entreprise Innovante). ● Write 3 Research Tax Credit reports with a detailed state-of-the-art and thorough description of all innovative works done by the start-up. ● Develop from scratch a Web application in Scala dedicated to the management of psychologist patients for the Paris area (Qualisocial). ● Supervise one student on Javascript development.

Many works have studied the Internet topology, but few have investigated the question of how it evolves over time. This paper focuses on the Internet routing IP-level topology and proposes a first step towards realistic modeling of its dynamics. We study periodic measurements of routing trees from a single monitor to a fixed destination set and identify invariant properties of its dynamics. Based on those observations, we then propose a model for the underlying mechanisms of the topology dynamics. Our model remains simple as it only incorporates load-balancing phenomena and routing changes. By extensive simulations, we show that, despite its simplicity, this model effectively captures the observed behaviors, thus providing key insights on relevant mechanisms governing the Internet routing dynamics. Besides, by confronting simulations over different kinds of topology, we also provide insights on which structural properties play a key role to explain the properties of the observed dynamics, which therefore strengthens the relevance of our model. ● Analyze a collection of IP routes collected at different series of times from the Internet. ● Propose novel statistical approaches to model the dynamic of the Internet topology. ● Design and develop a tool, Radarsimulator, to simulate the evolution of the Internet IP topology over time. Radarsimulator performs routing measurements on artificial graphs and generates different views of the graphs in several steps. ● Supervise 2 interns of undergraduate and engineer levels. ● 3 publications on national and International conferences.

My thesis was on developing more sophisticated trouble ticketing systems to ease network management. As a first step, I use information retrieval techniques to automatically analyze trouble tickets. This characterization allows network operators to quickly get statistics on past network incidents and can assist the troubleshooting of new incidents. Then, I analyze the impact of trouble tickets on both intra- and inter-domain routing. ● Build a tool to automatically and periodically scrap trouble tickets from various network operators. ● Design and develop an original tool, Troubleminer, that automatically classifies a collection of trouble tickets from different network operators. Troubleminer uses NLP (Natural Language Processing) and machine learning techniques to select relevant words on text documents, group them into clusters and structure them into a hierarchy. ● Collect and use network alarms to characterize the influence of different types of failures. ● Characterize the influence of different types of failures based on network alarms. ● Propose and implement different machine learning models to classify and predict network failures based on network alarms. ● 3 publications on International conferences.