Focus on expanding the data science, actuarial, and software requirements of BirdsEyeView. Main Tasks========Risk modelling for weather derivative contracts: Monte Carlo simulation, non-parametric kernel regression, extreme value analysis, stochastic calculus. Back-end development with focus on pricing services for real-time valuations of weather derivatives / insurance products.Data engineering, focussing on weather data ingestion and transformation pipelines.Communicating with brokers and underwriters for efficient delivery of new products, as well as investors.Backend=======Python + its data science ecosystem (pandas, numpy, numba, statsmodels, scikit-learn, ...)Java, Spring Boot, Maven, Docker for microservicesAWS for cloud infrastructure and services: S3, EMR, Athena, Glue, Lambda, EC2, ECR, EKSPostgreSQLFrontend=======Vue.jsManagement==========SlackJiraConfluence

Data scientist and engineer in an InsurTech startup company, building a weather risk insurance product.Main responsibilities:- Assessing, extracting and cleansing raw geospatial data from several sources to build data pipeline.- Designing predictive models to evaluate risk.- Identifying solutions compatible with business goals and software requirements.- Communicating with underwriters.

- Working with a team based at Imperial College London, UK, and groups world-wide. We meet virtually in CERN meeting rooms on a weekly basis, and present status updates and action plans.-Working with large and noisy datasets of the CMS experiment, searching for CP violation in the Higgs sector through the decay to a tau-lepton pair.- Developed and studied machine learning methods to categorise events into signal and background classes.- Optimised selections to maximise signal efficiency.- Simulated particle physics events to model the data, and derived corrections to ensure modelling is reasonable throughout the phase-space of interest.- Developing code in Python and C++ frameworks.

Worked on a cavity optomechanics experiment:- Characterising the motion of nanospheres inside the noisyenvironment of the hybrid electro-optical trap used in this specific experiment to gain a deeper theoretical and experimental understanding.- Conducting measurements and analysing data using the software programs PicoScope and Matlab to extract the Q- factor of the system.- Helping in the overall advancement of the experimental project by working in a team, suggesting ideas and assisting in some manual optical table work.

Studied and developed a simple genetic algorithm using Matlab and LabVIEW for the potential implementation in an experiment involving the detection and control of Rydberg atoms. The genetic algorithm was designed to solve a simple optimisation problem.

Worked in a team to: - improve and optimise the set-up of a magnetometer experiment, which involved technical work with optics and electronics, and understanding and application of quantum mechanics. - employ LabVIEW data acquisition tools to obtain the project's first real-time images of small metal objects. (See publication)

Gained detailed insight into ESOC's projects through lengthy one-to-one discussions with some of leaders of the operations centre.