Data Engineer and Developer in Višegrad, Republika Srpska, Bosnia and Herzegovina

Independently created and published a Udemy online course that's been taken by 23,000+ students. The course content teaches some foundational algorithms and computer science concepts and explains the principles behind fundamental algorithms asked about during interviews with top tech companies.

My team built a service to recommend the users who have the highest probability to complete a given survey. We were given a relational model dataset, which we transformed, enriched, and inserted into a PostgreSQL database. Then we built an ETL job that denormalizes the relational data model and extracts user-related aggregations into a NoSQL DynamoDB table used by our recommendation engine. We also created a recommendation pipeline orchestrated with AWS Step Functions to extract the most relevant user pool.

A distributed cloud-based analysis system that consumed and processed telecom data to create interactive data visualization. The tool included personalized package recommendations, which used clustering and a tool for detecting irregularities in the network in real time. The telecom data was stored in Elasticsearch, and Kibana was used for data visualizations and machine learning jobs to detect anomalies.

This project was the capstone project in the Udacity Data Engineering Nanodegree program. The idea was to create an optimized data lake that would enable users to analyze US accident data and determine the root causes of accidents. The main goal was to build an end-to-end data pipeline capable of processing big volumes of data. We wanted to clean, transform, and load the data to our optimized data lake on AWS S3. The data lake would consist of logical tables partitioned by certain columns to optimize query latency.

This project was part of the Udacity Data Engineering Nanodegree program. The task was to build an ETL pipeline that extracted data from AWS S3, staged it in Redshift, and transformed it into a set of dimensional tables that were ready analytics for consumption and enabled users to find insights into which songs certain users listened to. The output was a relational star schema in Redshift.

This project was part of the Udacity Data Engineering Nanodegree. The task was to take data from a source system and move it from a data warehouse to a data lake. The data resided in AWS S3, a directory of JSON logs of user activity on the app, and a directory with JSON metadata on the songs users listened to in the imaginary app.

The requirements included building an ETL pipeline that extracted user data from AWS S3, processing it using Spark, and loading the data back into AWS S3 as a set of dimensional tables. This would allow the analytics team to continue finding insights into what songs their users were listening to.

This project was part of the Udacity Data Engineering Nanodegree program. The task was to implement an automated and monitored ETL pipeline for loading data to a data warehouse. The pipeline was implemented using Apache Airflow. The source data resided in AWS S3 and needed to be processed in the target data warehouse in Amazon Redshift. The source datasets consisted of JSON logs of user activity in the application and JSON metadata about the songs the users listened to.