Ciklum is a global digital solutions company headquartered in London, UK.
19 мая 2015

Data Analyst for LoopMe (вакансия неактивна)

Днепр

On behalf of LoopMe (loopme.com/, Ciklum is looking for Data Analyst to join Dnepropetrovsk team on full-time basis.

A full time Data Scientist position for an exceptional candidate with graduate or doctorate in a mathematical discipline who has a keen interest in multivariate modelling, machine learning and decision automation.
This is a rare opportunity to be involved in creating an automated system to optimize hundreds of real-time interactions per second with online customers, and to quantify the benefits created against control groups in a live environment. The work involves analyzing billions of records of ad-serving opportunities, ad-serves and ad-serve responses to understand what factors explain the variance in expected response value. That understanding is then to be harnessed as an automated optimization process capable of intelligently finding and servicing the most favorable ad-serving opportunities in a way that leads to higher returns on our clients’ advertising investments. Guidance is available from a senior analytical party who will support the work.
This is also an opportunity to become part of the UK tech start-up world and get first-hand experience into how tech start-ups operate.

Responsibilities:
• Define ad delivery optimization tasks based on your data analysis
• Maintain self-learning optimization algorithms
• Write product specs for ad delivery optimization specs
• Define optimized audience segment queries

Requirements:
• At least three years practical experience of univariate and multivariate statistical analysis in ’R’ with large data sets (millions of records and many tens or hundreds of independent variables)
• At least two years of experience with building robust multivariate predictive models in ’R’, and the validation, application and monitoring of such models
• Good experience with SQL for the extraction, preparation, manipulation of data sets for the purposes of analytical preprocessing
• Good experience of variable transformation and data preprocessing techniques to extract maximum predictive power such as binning, piecewise linear regression, non-linear function transforms, etc.
• Good experience across a range of variable types: discrete and continuous, ordinal and nominal
• Excellent practical knowledge of multivariate techniques such as: Clustering, GLM, Logistic Regression, Decision Trees, Random Forest, Naïve Bayes, B&B, etc. and a good grasp
• SQL and NoSQL database literacy — proven skills to support the extraction, preparation, manipulation of large (multi-million record) data sets for the purposes of analytical preprocessing
• ’R’ (www.r-project.org) — excellent skills in programming and working in ’R’ to support the following steps in the analytical process with large (multi-million record) data sets: Basic data cleansing and preparation; Variable preprocessing/transformation; Univariate analysis; Performing statistical tests; Covariance analysis; Generation of graphical output; Multivariate analysis and linear/non-linear regression; Preparation of data sets for predictive modelling; Robust predictive model building, validation and application; Automation of statistical processes
• Proven ability to simplify statistical information, extract and present what really matters to less statistically knowledgeable peers and management
• A minimum of a Bachelor’s or Master’s degree in a mathematical discipline such as Applied Statistics, Maths, Engineering or Physics from a respected University
• Possibly a PhD in a mathematical discipline, though not a requirement

Personal skills:
• You love mobile
• You are highly intelligent and articulate
• You demonstrate a high level of initiative
• You have an enquiring mind and a disciplined scientific approach to extracting facts and understanding observed behavior
• You are excited by the potential of analytical intelligence to realize high-value commercial outcomes and change the way that business operates
• You consistently deliver high quality answers
• You want to be part of a high growth startup company with global ambitions

Desirable:
• Experience with statistical analysis in a marketing or advertising context with binary dependent variables that have low response rates (RR <0.1% say)
• Familiarity with stratified sampling
• Experience with the design and analysis of A/B split tests
• Experience with the design of multi-factor tests and efficient multifactor testing methods such as Taguchi methods
• Experience in the productization or automation of an end-to-end statistical process (such as data extraction, preprocessing, analysis, predictive model-building, application, performance monitoring and model retirement/updating)

What’s in it for you?
• Competitive compensation package
• Hundreds of millions of people see your work and use our products worldwide
• A transparent work environment
• Working in a team that inspires, and works with you to achieve industry changing goals

Join Ciklum and “Cross the Borders” together with us!
If you are interested — please send your CV to bovl@ciklum.com