Ertuğrul Ural

Company: Eurowings is the Lufthansa Group’s low-cost airline and part of the world’s largest aviation group.

Responsibilities:

• Working with stakeholders to develop quarterly roadmaps based on impact, effort and test coordinations.
• Managing the full model lifecycle from data to production.Through MLflow tracking experiments of the team and controlling parameters, metrics of the models and versioning of data and code, as well as model artifacts with each training run.

Achievements:

• Dynamic Pricing: Machine learning based pricing recommendations for ancillary products specific to each customer interaction and optimizes expected revenue per customer. (9 million euro additional turnover)
• Destination Recommender: Based on user interaction with the data source (flight transactions and flight searches) the algorithm learns the behavioral patterns of the customer in the historic data, then generates personal destination recommendations for each customer. (Conversion rate increased by 25%)
• Ancillary Recommender:The recommendation engine provides a list of the most proper flight ancillary products based on the previous purchases of the customer (Revenue increased by 6%)
• Customer Matrix: Customer Data Mart and the business reason of that product is to get 360-degree view of customers.
• Booking Forecast Model: Time series model forecasting the bookings within 6 months at route level. It's used for capacity planning and steering the campaigns.

Company: Vodafone is a multinational telecommunications company and its global headquarter is based in London, England.

Responsibilities:

• Creating advanced analytics and machine learning solutions that provide actionable business insight to Vodafone and its stakeholders.
• Developing predictive models with large and varied datasets, working with a community of colleagues across advanced analytics, technology, and data and customer functions.

Achievements:

• Churn Model for In-contracted Customers: 10% increase in precision value for the top ten percentile predictions.

• Prediction of Network Related Calls: Prediction of network related calls to reduce staff requirements and cost per-call.

• NLP Projects: Identifying churn reasons by extracting churn related topics/keywords from Speech to Text Data, Customer Review Classification, Analyzing relation between Churn and TNPS.

Company: Tüpraş, Turkey’s sole oil refiner and Turkey’s largest industrial company by revenue and added-value generated.

Responsibilities:

• Originates and carries out engineering studies to determine potential problems and remedies to existing problems resolve complex technical problems
• Increase revenue, optimize processes, protect the environment, or improve safety.
• Helping Tüpraş to overcome digital transformation challenges.
  

Achievements:

• Development of Quality Estimator For Heavy Diesel: aims to develop quality estimator for heavy diesel, which is 95% end point of heavy diesel, drawn from atmospheric distillation column of the crude oil unit using the tree based machine learning algorithms. The project is implemented on Tüpraş İzmit Refinery and modelling data are collected from real process data and laboratory assays.

•Predictive Online Spalling: Online Spalling is an operation implemented in Delayed Coker Unit to remove coke deposits from the inside of the coker furnace tubes. The primary benefit of the project after finding the temperature patterns predicts the spalling period and in this way can maximize the time period for the furnace to stay in operation.

•Anomaly Detection: In a modern refinery, most all of the installed equipment include sensors for monitoring their behavior and to avoid undesired events and to prevent possible catastrophic failures. Anomaly detection is the problem of finding patterns in data that do not conform to an a priori expected behavior.

About Company

•Tanı is a CRM company, which run 29 different customer loyalty programs for 170 brands operating in 23 different sectors

Responsibilities

• Provide actionable and insightful recommendations from data science, for segmentation, customer profiling and predictive analyses with the goals of: identifying acquisition, cross-sell, up-sell, retention opportunities and predictive churn analysis.

Achievements

• Digital Insert Project :Recommendation Engine Project to offer right products to right customers ==> Increased campaign performance (approximately 4.5X Campaign Return Rate)
• Churn Projects for retaining valuable customers ==> Min. %10 Increased model accuracy
• AIG is one of Tanı's clients operating in insurance market, Propensity models for growing customer base ==> %20 Increased campaign return rate
• Opet is one of Tanı's clients operating in fuel distribution market, Customer Segmentation/Lifetime Value Project ==> To gain actionable insights for marketing actions
• Koçtaş is one of Tanı's clients operating in home-improvement sector, Basket Analysis ==> To uncover associations between items

Company: Türkiye İş Bankası or simply İsbank is Turkey's largest bank. It was the first public bank founded in the Turkey.

Responsibilities:

• Calculation of development index in cities and districts in order to capture potential of sales regions of bank branches through leveraging demographic data.
• Take initiative to research, identify, and deliver learning opportunities to optimize branch sales targets through parametric and non parametric models.

Achievements

• Increased operational efficiency %50, sales target accuracy %20.