Ashkan Khoshcheshmi

• Have been working on designing and creating cloud data warehouse (CDW), using Data Vault methodology, in Snowflake from ground up with goal of new system becoming hub and spoke model for company's data. Company's data mainly consists of ticketing, fundraising, and marketing data that powers more than 500 live entertainment organizations.
  Team size on this project has been varying from 5 to 12 people.
• Led research which resulted in company's choice of Snowflake for cloud data warehouse, Matillion for ETL tool, and Data Vault for data modeling methodology.
• Led design and implementation, for both DDL and ETL, of following layers in Snowflake: Data Lake -> Stage -> Data Vault -> Infomart (star schema).
• Wrote Python code for transforming data from data lake into Data Vault's rawvualt layer which dynamically creates transformation SQL statements based on meta data stored in control table.
• Collaborated on creating following two main data gateways into Snowflake where data is being ingested in semi-structured (JSON) format into data lake:
  - Streaming route using Confluent Kafka
  - File-based route using AWS S3 and Snowpipe
• Data at presentation layer (star schema) is being consumed by BI tools (currently and mainly Tableau, previously Looker).
• Collaborated on ETL (Extract, Transform, Load) tasks, maintaining data integrity and verifying pipeline stability.
• Designed and developed analytical data structures.
• Designed and implemented effective database solutions and models to store and retrieve data.
• Developed database architectural strategies at modeling, design and implementation stages to address business or industry requirements.

Environment: Snowflake, Data Vault (JSON-based Satellites), AWS, Visual Studio Code, PyCharm, Anaconda3, Flyway, Miro, GIT, Looker, Tableau, Jira Agile, Visual Paradigm (ERD), Confluent Kafka, SQL Server

• Worked on Data Lake project where semi-structured data was ingested onto cloud data warehouse to support course/system usage reporting, statistical analysis for predicting at-risk students in course, and market research/analysis. Team size on this project varied from 20 to 40 people.
• Student and course activity data from different sources were being streamed onto AWS S3 through Kafka clusters, transformed/aggregated in Snowflake cloud data warehouse, and eventually used for reporting and predictive analysis.
• Underlying process for reports that would originally take days to generate for course instructors and system administrators, was migrated to Snowflake which reduced report generation time to minutes. Developed provisioning, deployment, and configuration process that client would use to implement migration to this new pipeline.
• Goal for centralized data warehouse was to be able to dynamically scale on-demand for different systems serving 350+ and 1,400+ clients respectively.
• Created workflow/process for ingesting raw semi-structured course access logs and transforming into multiple relational tables on cloud that would support Looker reports which provided with charts and dashboards used by internal staff. Above-mentioned access log table was receiving 100+ million rows of raw data per day so it had to be archived as one of last ELT steps.
• Used GIT to collaborate with team members on Data Lake's codebase. Used Jira agile board for task tracking and visualizing team activity that were reported to product management at end of each sprint.
• Different tools and languages used as part of Data Lake project: Hue, Hive, Pig, Impala, Presto, Sqoop, Oozie, Flyway, AWS (S3, EC2, EMR, RDS), jq for JSON, IntelliJ, Shell script, Scala, JavaScript, Node.js, Highcharts, Snowflake, Looker, Python, PyCharm, Airflow

Environment: Cloudera CDH, AWS, Stash (GIT), Jira Agile, SQL Server 2005/2008/2012, SSRS, SSIS, BIDS Helper 2012, ASP.NET, Visual C#, Windows Server 2008 R2 / Vista Enterprise, MS Visual Studio 2013, MS Visual Web Developer 2008, SQL Server Data Tools, MS Visio

• Managed Analysis Services and Reporting environment of Blackboard Connect, subsidiary of Blackboard where 6+TB of relational/multidimensional data warehouse was being developed and maintained. As result, data analyses and reporting activities were offloaded from transactional databases and platform for different BI needs was provided.
• Setup of complete Business Intelligence environments that would provide with various BI capabilities such as reports, dashboards, and online analytical processing (OLAP) where slicing & dicing of data in number of different ways would take place.
• Created and maintained reports from relational and/or multidimensional databases. These reports would provide with visibility into core activities of company's main products and were being utilized by both clients as well as internal stakeholders. Combination of T-SQL, SSIS package transformations, and MDX queries were being used in this process.
• Setup of nightly cube process job where data was pulled from approximately 70 different OLTP tables and subsequently was inserted into OLAP in incremental fashion. This process was keeping data up-to-date in both relational data warehouse as well as OLAP environment.
• Used TFS (Team Foundation Server) in Visual Studio for version control, work item and bug tracking, and to coordinate work with other team members.