Howard Cho

As the only Data Engineer at Deepgram, I have built out a lot in a very short amount of time. Without a team of data engineers helping, I needed to make decisions that allowed me to get up-to-speed quickly, which is why some tools were chosen (ex: Glue vs Spark and/or Flink). * Built out a transactional Data Lakehouse using Apache Hudi with a Medallion architecture pattern * Setup replication for the Bronze-level data from a variety of sources including NATS JetStream, Postgres (Aurora), Salesforce, etc. * Implemented CDC (Change Data Capture) from PostgreSQL to our Data Lakehouse utilizing Debezium Server, and writing to NATS JetStream * Built micro-batched Salesforce exports from NATS JetStream, removing reliance on a 3rd party data management tool, saving ~$50k, annually * Tasked with getting row counts down in our main BI tool from over 200M rows to under 15M to avoid a surcharge; completed the work in under 3 months, saving an estimated $30k jump in annual costs * Completely removed reliance on our main BI tool in favor of a hosted Superset instance, saving ~$70k, annually. * Led research into a replacement visualization tool for our main BI tool, ultimately landing on Apache Superset * Created multiple Gold-level datasets for use in Superset, allowing the Product team to cutover from our main BI tool * Automated hourly aggregated data for our secondary BI tool ingestion, Rill

* Managed the Data Platform team, and grew the team from 3 to 6 Data Engineers * Prepared documentation, solicited feedback, and successfully promoted two Data Engineers to Senior level * Targeting a new latency of under an hour with a 4x increase in data volume, rallied the team to refresh the event streaming ETL in under a week, and successfully reduced latency to 30 minutes into Snowflake for all (~100) streaming event types; this resulted in no latency during the Black Friday/Cyber Monday weekend (highest volume weekend) * Planned, prioritized, and prepared a roadmap to adhere to a new 15 minute SLO for all streaming events into Snowflake; completed the development within 3 months (on-time) with full validation, ensuring no data loss, ultimately leading to no latency for a subsequent Black Friday weekend * Designed and implemented a new stream ingestion process utilizing Apache Pulsar and Flink processing to reduce event ingestion latency with multiple outputs including Snowflake, Druid, and our Data Lake; this included validation processes using Spark * Successfully developed a Junior engineer into a highly performant team member who led two major, multi-team initiatives as a tech lead * In June 2022, expanded role to include managing the Reporting Platform team, growing the team from 4 to 5 engineers, consisting of 3 Front-End/Full-Stack engineers and 2 Back-End engineers; the team also included, 1 designer, and 1 PM * Inheriting two projects past deadline, worked with PM and stakeholders to reprioritize the projects based on team availability and impact, ultimately delivering both projects successfully and removing the need for manual work from our CS team * Facing delays of ~9 months on multiple reporting projects (inherited), worked with stakeholders to determine the highest impact items of each of the projects, broke the project into multiple milestones, and successfully released the highest impact items, driving significantly increased usage

Fat Buddha Bar is a Korean Hip Hop bar/restaurant/event space that is located in the East Village, NYC. We remixed a failing business by streamlining the kitchen menu, rebranded, designed and built a new layout, and curated great music/events. We have the best staff, and are extremely proud of the whole team.

* Built a replication of the monolithic MySQL database’s data into Snowflake within the first two weeks of starting * Designed and built the initial ETL of streaming events into Snowflake utilizing Kinesis, AWS Lambda, AWS S3, and Python on an hourly cadence * Defined and documented the “DB Best Practices” style guide, and planned and recorded a video training on SQL Querying Best Practices for the wider engineering organization * Developed the interview questions and process for Data Engineering, successfully used for the past >2 years

As a full-stack developer, I've worked on all aspects of our primary products (DotLook & DotAlign Cloud) using a series of languages and technologies including: Azure Cloud Computing, SQL Server, C#, F#, PowerShell ISE, C++, Elm, Kubernetes, Docker, and SQLite. With my strengths in database querying, I also spent a lot of time refactoring and optimizing existing queries to increase performance.

Focused on all aspects of the business, including operations, strategy, and development. As the lead database and infrastructure developer, I handled the build-out of the servers utilizing Amazon Web Service's EC2 and RDS instances, running Ubuntu. Additionally, deployed and administered the MySQL and Solr databases, along with managing heavy data load requirements, primarily utilizing Java. Relied on to handle significant development work of the middle-tier, API-driven web services in Java, utilizing Spring, Maven, c3p0, and Git for source control. Planned and built the primary application from concept to full integration in under 1.5 years for a single commodity, and added a second commodity in under 2 years. The software had become the primary application at a leading Energy Supply Company, requiring less than $600k in investment, while generating over $240k, annually.

As the only DBA here, I am tasked with doing pretty much anything database related. Projects Completed: - Migration of databases from the previous hosting solution to a Rackspace environment - Redesigned database replication schema from a star design to a more highly-available dual-master replication scenario, with each master having an additional read-only slave. - Setup, tested, and implemented MySQL Multi-Master Manager (MMM) to manage reader/writer VIPs, enabling server maintenance with zero downtime, along with auto-failover of the primary master. - Implemented Percona tools to ensure database consistency.

DealMaven was purchased by FactSet in February 2008. Managed a team of two data engineers. Accountable for all aspects of database administration, architecture, and development for a series of new products and added functionality of existing products primarily using MySQL. Tasked with planning, developing, and maintaining applications to deal with all aspects of data handling including (but not limited to) data manipulation, data shipping, up-time, etc, in C#.NET, Java, and Perl. Consult with differing departments to integrate data sets efficiently and effectively while minimizing development time.

Accountable for analyzing, planning, designing, and managing the utilization of MS SQL Server to handle existing and newly added functionality to the company's online testing and web presence. Worked with a team of developers to create and interface with differing data sets, allowing for more rich and robust information to be provided to end users.

Initially hired as an independent contractor to migrate data from a home-built database to a canned solution (Raiser's Edge) using a Sybase database, and to train the staff on how to use the software. Contract renewed and extended to include all aspects of I/T support including network administration, hardware/software support, database administration and development, and custom software developer (VB6).