Sergey Zhiganov

Программист @Deco Systems

Moscow, RU

More (RU): of classification and response systems- System for primary classification of requests: Development using PosgreSQL, Airflow, CatBoost, BERT stacks- Initial response system for HR requests: OpenChat 3.5, FAISS, Flask, BERT, LongChain, with visualization in Superset.Development of a corporate data warehouse system (Python, Airflow, Spark)- Greenplum assembly- Development of Ansible roles for Greenplum installation- Implementation of data delivery modules- Development of a process for updating storage layers- Introduction of the DV 2.0 concept- Writing template update scripts with versioning of directories.Migration of Hadoop 2.6.5 to Hadoop 3.0.1 (Ansible, Python)- Migration of the source code of streaming projects to new versions of Kafka, Spark, Flink- Development of Ansible roles to automate installation and management with support for Kerberos and HA- Correction of Spark/Flink application code and adaptation of Hive SQL syntax- Analysis and elimination of calculation errors, identifying the causes of differences.Development of a data masking system (Python, Airflow, Spark)- Support for pluggable Python code at runtime to process variables and results- Support for dynamic generation of UDF functions for data masking.Development of a data quality system (Python, Airflow, Spark.)- Dynamic formation of a graph of the verification process based on metadata- A mechanism for generating and recalculating calculated variables with saving the results- Execution of plug-in Python code.Development of a transaction analysis system (Python, Airflow, ML algorithms.)- Preprocessing of input parameters- Implementation of a module for clustering transactions using ML algorithms- Construction of metadata for visualization of graphical representation of results.

ФГБОУ ВО КнАГУ

Аспирант

Комсомольский-На-Амуре Государственный Технический Университет

Кандидат Техни

ФГБОУ ВО КнАГУ

Магистр

ФГБОУ ВО АмГПГУ

Бакалавр