ML response prediction and fraud detection. I worked as horizoom's first ML Engineer within the Tech & Product team. I built a response prediction model to predict whether panelists will respond to survey invites within different timeframes. I collaborated on the feature engineering with a topic expert and BI Analyst, fine-tuned the model on custom performance metrics and deployed both the training and prediction pipelines in Airflow, closely collaborating with Data Engineers.I also developed a clustering-based approach to identify groups of fraudulent accounts. I liaised with Customer Support and Product teams to agree on the most critical types of fraud that should be targeted and focused my work accordingly. In the end, I built an interactive dashboard populated by my results that allows stakeholders to block fraudsters effectively.

Statistics & Experimental Methods

I was responsible for all analytics needs of two Product teams, closely collaborating with Product and Engineering Managers. In my team, we ran monthly ML hackathons, prototyping new predictive models for various parts of the product and its architecture. Examples included appointment acceptance prediction and event outage modelling. General responsibilities included: - Exploratory analysis projects to find useful patterns in user behaviour - Prototyping of new ML models - Designing experiments to test new product features - Development and calculation of KPI equations - Event tracking and dashboard maintenance and improvements

I worked in Blinkist's BI team on all problems related to product and user behaviour. I also led a weekly meditation session. - Event tracking- KPI definition & calculation- Amplitude instrumentation & training- Matillion ETL setup & maintenance- A/B test ideation & analysis- SQL (Redshift), Python, Google Script, Git

I travelled through Thailand, Myanmar, Malaysia, and Singapore. In this time, I attended two silent meditation retreats (10 and 7 days long) to deepen my mindfulness practice and develop greater understanding of Buddhist philosophy.

Behavioural Design & Big Data Analytics:I used my understanding of Behavioural Science to inform the design and operations of places with large crowds of people, primarily in the transport sector. I have collaborated with experts from both the industry and academia. I analysed and visualised large datasets, such as cellular mobile phone data, Bluetooth beacon data, as well as more traditional sources like survey or administrative data, to derive insights into people movement, both on a macro level (such as a countrywide information system for highway traffic) and on a micro level (such as visitor movement patterns in a museum). On the technical side I used a healthy combination of R, SQL, Python and Excel. Having done consultancy work for 2.5 years, I am used to working against tight deadlines, handling the demands of multiple projects, and presenting results to clients in a clear and concise way. I have mentored new starters, and also designed and delivered technical training courses to get team members up to speed with our work.

I was in charge of data analysis in an interdisciplinary research project on consumer behavior. This project, run in collaboration with a major UK retailer, took a data-driven approach to investigate how consumers make decisions in a real-life shopping context. I also ran a social media analytics project on data from the Twitter API.

I was employed in the Methodology section of the Institute of Psychology. I lead two weekly lab classes for undergraduate students on the contents of the Statistics class. I was responsible for the learning progress of ~60 students each semester, so my main task was to help students understand the topics discussed in the Statistics lectures and apply it in weekly homework assignments. To this end, I extracted the essential parts from the lecture scripts into concise summaries with suitable examples for each topic. In addition, I occasionally drafted presentations to illustrate particularly complex problems and help the students gain further insights. Moreover, I individually supported and prepared undergraduate students for their Statistics exams. Finally I was involved in the grading of the exams upon the end of each term. Throughout this time, I greatly improved my self-reliance and communication skills. In creating summaries, slides and small programs I also gained analytical and computational skills.I carried out all analyses and graphical illustrations in R.

My responsibilities included planning, conducting and evaluating empirical research, run in the Methodology department by Prof. Wolf Schwarz.In collaboration with two other research assistants I strengthened my teamwork skills in organising research projects. In addition, I gained valuable insights into the practicalities of computer-based empirical research.

I tutored high school students, mainly in Latin, also in Maths and English.