Nikola Chochkov

Responsible for solving Flowkey’s long-term strategic challenges in the area of Data Engineering and Analytics:

• Data Warehouse Design and Implementation
• ETL processes
• Business Intelligence/Reporting/Growth Accounting
• Statistical Data Analysis
• Development, Deployment & Monitoring of Data Science products
• Project Management/Training/Team Management

Technologies:

• PostgreSQL/PLpgSQL/Timescale
• Python/R
• pandas/data.table
• Apache Airflow
• Metabase/Jupyter/Superset/Rmarkdown
• 3rd Party Integration (Mixpanel, Adjust, Facebook Marketing API, Chartmogul, etc.)
• AWS Services

I provide consulting services and build/integrate solutions for my customers in the areas of:

• Data Warehouse Design and Implementation
• ETL processes
• Business Intelligence/Reporting/Growth Accounting
• Statistical Data Analysis
• Development, Deployment & Monitoring of Data Science products
• Project Management/Training/Team Management

Portfolio:

• flowkey.com - Data Warehouse Design, ETL, BI Reporting and Visualisation, Statistical Data Analysis
• fatmap.com - Data Warehouse Design, ETL, BI Reporting and Visualisation, Statistical Data Analysis
• viaeurope.com - Analytics Database, Query Optimisation, ETL, BI Visualisation

Technologies:

• PostgreSQL/PLpgSQL
• Python/R
• pandas/data.table
• Apache Airflow
• Metabase/Jupyter/Superset
• 3rd Party Integration (Mixpanel, Adjust, Facebook Marketing API, Chartmogul, etc.)
• AWS

Data Science challenges grew at Adjust in the areas of Anti Fraud, Business Intelligence, Anomaly Detection, Marketing and others. I was in charge of hiring and leading the Data Science team, which within the first year:

• Built and deployed a custom Anomaly Detection system, monitoring our customers’ apps data streams and alerting Customer Success team on suspicious patterns.
• Deployed a company-wide (more than 100 users) Metabase instance, currently serving as an interface to nearly all of Adjust’s internal data analytics needs.
• Built and maintained the computation pipelines serving product features like www.app-benchmarks.adjust.com
• Built data science models for user-level churn-prediction, lead-scoring and others
• Built automations for analysis of large, multidimensional datasets

Involvement:

• Overseeing the development of all data science products during all phases from statistical modelling to deployment
• Coordinated the company-wide cross-team effort to adopt Metabase
• Hiring and training the data science team

Technologies:

• R, Python, PostgreSQL, Flink, Metabase, R Shiny, Apache Airflow, UNIX automation
• Clustering, XGBoost, various Anomaly Detection techniques

By mid 2016 Adjust had reached an incoming traffic rate of 100K HTTP requests per second and our tracking SDK was already estimated to be found in every second mobile device in the world. We had two clusters of data - raw and aggregated, each consisting of doezens of PostgreSQL servers. We had introduced Flink and Kafka to our technology stack.

I joined the growing Database team contributing by maintaining and developing PostgreSQL extensions and building the Analytics toolsets for our sharded environment.

Technologies:

• PostgreSQL, Flink, Kafka, C, R

I joined Adjust at a very early stage as the company’s 5th employee. Over the following years, Adjust experienced a rapid growth in scale and team members and I contributed in multiple roles, everywhere where the company expansion required it.

Involvement:

• Building analytics features of the product such as Sentiment Analysis on Mobile App Reviews, the GlobalRank algorithm for the iTunes and PlayStore apps and others
• Numerous Data Analysis projects involving big and rich datasets of app-tracking (installs, events, sessions) and app-store (rankings, reviews) data
• Development of microservices, based Ruby on Rails and Golang
• PostgreSQL development and optimisation

Technologies:

• R, PostgreSQL, Ruby on Rails, Golang, C, Microservices, 3rd party APIs

As a freelance contractor I analysed my customer’s web-scrapped data and developed Regression-based software for price-prediction of used cars, properties and boats. The software is still in use by insurance customers of my employer.

Technologies:

• R, PostgreSQL, GLM models

I analysed data from SoundCloud comments, posts and group messages to develop an ML prototype that achieved 97% accuracy.

Technologies:

• Ruby, Bayesian Inference

During that period I had numerous contracts for web development for startups in Berlin and outside. Most significant project was at digitaleseiten.de building the Ruby on Rails platform supporting the company’s product websites.

Technologies:

• Ruby on Rails, Php, PostgreSQL, MySQL, Javascript, etc.