Aleksandr Sharifullin Email & Phone Number

Amsterdam, North Holland, Netherlands

- Modernizing current data infrastructure using up to date data manipulation tool set (Airflow, Pyspark, Snowflake, dbt)- Optimizing data delivery time- Adjusting data ecosystem in terms of law compliance

✓ Implemented records deduplication in a large-scale data pipeline in a distributed Hadoop-like environment for obtaining AI datasets which decreased train data preparation time and resource consumption x2 times✓ Reduced overall CPU, HDD consumption by 10% tuning SQL, Spark tasks performance in large-scale ETL pipeline✓ Developed data quality check system on Python with monitoring based on Datalens for AI data processing large-scale MapReduce ETL/ELT-pipeline which decreased time for trouble fixing to hours except days✓ Developed Grafana dashboard for computing resource usage monitoring by data processing tasks (CPU, RAM, etc)✓ Designed a solution for predicting resource consumption by new upcoming tasks using statistics from production✓ Invented a culture of documentation for data entities and designed unified standards of computer-readable documentation, thus decreasing the time for onboarding new people in the company✓ Communicated with data providers and consumers to improve data model details inside the AI data pipeline✓ Reorganized project management in our data engineering team (x4 faster task burn rate increased)

✓ Developed ETL-pipelines based on Apache NiFi, Apache Airflow, also SQL+Unix Shell decreasing data refreshing time by over 5 times and maintained them in Docker-based production✓ Created the bootstrap-based SQL solution for fast initialization/rollback of universal data vault based on PostgreSQL DBMS (x4 speed up in DWH building)✓ Developed a WEB-based solution using python+spark+SQL for predicting the operating time of a thermophysical installation based on readings from Kafka with Docker containers deployment in K8S orchestration✓ Communicated findings and insights to stakeholders, worked with cross-functional teams to gather requirements and design data models

✓ Implemented a dispersion analysis to get information about diversity between lithology classes in 4-dimensional space✓ Constructed neural networks with different architectures (feed-forward, recurrent) to compare their inference in point-by-point class prediction in data which allowed estimation of geologic layer-to-layer dependency✓ Developed a library for comfortable well-logging curves visualization✓ The results of the work are published in the journal "Georesources": https://geors.ru/archive/ article/1166/

✓ Developed an algorithm of well bore repair duration definition according to device readings✓ Implemented specific data preprocessing including low-frequency filtering, MA smoothing, and local trend line estimation in a automatic pipeline✓ Trained anomaly detection isolating forests algorithm for the determination of repair process start/end

Master'S Degree, Artificial Intelligence, Excellent

Activities and Societies: I participated in many AI/ML hackathons and kaggle-like competitions. Also, I participated in Sirius schools.

Bachelor'S Degree, Geophysics And Seismology, Excellent

Activities and Societies: I was a member of the Student Council and an actor in a faculty theatre during summer geology practices. Also, I.