Arthur Krieger

• ›Designed and built a declarative metadata reconciliation system between mid April and the end of August 2026, taking it from the first conversation to production. It stops schema deployments from silently removing Snowflake data governance metadata, and it cut recovery of thousands of metadata rows from three to four hours of manual rework to seconds
• ›Built it to be safe to re-run, so a failure leaves data as it was instead of corrupting it, and recovery is a replay rather than a rebuild. It has handled 20+ schema deployments since go-live without anyone stepping in
• ›Made it work across warehouse object types and both standard and custom metadata, with each team onboarding itself and managing its own metadata. Non-engineers edit metadata in the spreadsheets they already use, without writing YAML or waiting on an engineer
• ›Build and maintain data pipelines that ingest, transform, and deliver financial data to analysts and data scientists on the Wealth Asset Management Engineering Team
• ›Develop GitHub Actions CI/CD workflows and DevOps automation for the engineering team
• ›Create Streamlit dashboards for internal data observability, pipeline monitoring, and ad-hoc analytics on Snowflake
• ›Provision and manage cloud infrastructure with Terraform on GCP

• ›Part of a scrum team using Agile methodology. Built dashboards for data observability and monitoring of data pipelines using Streamlit and Snowflake
• ›Delivered Streamlit dashboards displaying pipeline operations and data profiling metrics using Snowflake
• ›Presented a research plan on using generative AI to synthesize data and improve fraud protection

• ›Integrated Azure DevOps REST API in Databricks using Python and SQL to track software changes
• ›Researched and implemented External Tables in Snowflake to reduce costs and bypass unnecessary data transfers
• ›Presented to stakeholders at the end of each 3-week sprint

• ›Used R to analyze gene expression rates and auto generate visualizations for a published preeclampsia research paper
• ›Built an R Shiny app for researchers to query an SQL database and generate plots interactively

A personal platform I build and use for my own apps, across 18 repositories. The apps share single sign on and single logout through an OAuth 2.0 and OIDC identity provider, common code lives in shared Python and npm packages, and Terraform and a CI/CD library keep every repository built and deployed the same way.

A boutique storefront I am building for my sister, where she will sell curated clothing, books, handmade crafts, art, and accessories. The server works out every price, total, and stock count itself, and checkout takes two steps, reserving the items and then paying.

A simple food and nutrition tracker for casual users and beginners. It logs meals and tracks calories and macros against daily goals, with a food database that includes about a thousand foods from USDA data.