Andrei Loshchev

 Hello, I'm Andrei!
I have spent the last years doing study and work on various Data Science projects in Business and Social industries
with different kind of software and I am keen to continue developing my career in the field of Data Science, Analysis,
BI and AI development.
Some notable examples of my career story include:
-Worked on a great Data Science projects such as KAI (realtime AI for Ad Fraud prevention) where I developed
Unsupervised and Supervised solutions to detect various botnets and block up to 30% fraudulent traffic and ML text and table detection tool for N-SCR reports analysis with 90% accuracy.
-I have developed and deployed a BI web-service for analysis and forecast POS Time Series Data.
On a personal level, I can describe myself as a very calm, stress-resistant, friendly, organized and responsible person in my work and life. I have good communication skills and like to make jokes, share and present my ideas. I can work with minimum supervision and pick up new knowledge very fast, but at the same time I enjoy being part of a team of different people, I always ready to help and bring my own point of view and creativity. I’m also  really into sport, creative activities, traveling, and learning foreign languages and explore new cultures.

Data Scientist Kubient, New York City Product: • Developed a solution that blocks fraud before it happens. The ANN-based model blocks 14 different types of fraud (Datacenter IP fraud, Botnet activity, Bundle spoofing, Entropy fraud, etc.) in real time Responsibilities: • Raw data processing and analytics • Data analysis and anomaly detection in bid requests from Google BigTable • Development of Supervised and Unsupervised machine learning models for real-time fraud detection • Creation of an ETL process using Big Data services • Training and testing machine learning models • Run trained models in real time • Analysis of all types of internet traffic (websites, mobile applications, CTV) Achievements: • Created a model with an average response speed of up to 10 ms • Product based on real-time machine learning solutions successfully blocked 92% of fraudulent transactions • Fraudulent traffic was found to account for up to 30% of total customer traffic • Successful detection of large-scale botnets

Graduate Diploma in IT

Otago Polytechnic - Auckland • Relevant papers: Data Science&Machine Intelligence, Databases, System Analysis, Webprogramming • Awarded Data Science&Machine Intelligence (Excellent Scholars Award OPAIC