A bit more about me

Our Dremio data platform: a production service at three nines that backs a customer-facing product and about a dozen internal teams. One query engine over 30+ data sources, plus the streaming and AI tooling around it.

• Took the Dremio lakehouse from a single-VM proof of concept to a production-critical service at three-nines availability that backs a customer-facing product and about a dozen internal teams: dev, staging, and prod on Kubernetes from a maintained Helm chart, deployed through GitLab CI/CD, behind F5 ingress, with RBAC and vault-synced secrets.
• Federated 30+ data sources across six backend types (HDFS, Hive, S3, Snowflake, Postgres, Mongo) behind one SQL engine with an Apache Iceberg catalog and semantic layer, joinable in a single query.
• Cut multi-minute analytical queries to seconds by isolating heavy workloads onto dedicated reserved engines and tuning query parallelism on partition-heavy datasets.
• Own on-call, incident response, and the HA/DR posture (RTO/RPO targets, coordinator HA, cross-region backups). Resolved a production storage-auth outage in about 90 minutes, then used the window to remove the cloud-credential dependency that caused it rather than just patch it.
• Drove the build-vs-buy decision for a streaming platform through a day-2-operations bake-off, then built it on Confluent for Kubernetes (KRaft, OAuth/RBAC, TLS, schema registry, Connect) with Debezium CDC landing topics in the lakehouse as Iceberg tables.
• Built an AI data steward: a two-phase LLM pipeline that documents the catalog under a cost guardrail, an MCP server engineers query from their editor, and bronze/gold observability over its runs.
• Own SSO and OIDC for Dremio and Superset, including migrating the platform identity provider, partnering with the identity and security teams on access and group mapping.
• Run Apache Superset as the BI layer over the lakehouse, plus a one-click generator that turns any dataset into a starter dashboard.
• Primary technical liaison to DevOps, networking, and cloud-security teams; integrated the first Prometheus/Grafana/JMX metrics and query/audit logging for the platform.

Worked on the company's Data Mesh platform supporting governed data access, analytics enablement, and internal data products.

• Stood up the team's first Dremio lakehouse on Kubernetes (an Iceberg catalog, Kerberized HDFS in HA, and federated on-prem and cloud sources): the foundation the production platform grew from.
• Built the platform's first data catalog, a nightly crawler that loaded ~4,000 data assets with schemas, lineage, and owners, and made it boring and fast: import errors dropped from ~700 to ~30 per run and runtime from ~34h to ~13h.
• Designed the data-mesh API layer and built MuleSoft proxy APIs over internal RPC services, collapsing a three-step auth flow into one authenticated REST call across dev, QA, and prod.
• Built and launched the team's internal Next.js portal (SSO, Kubernetes, CI/CD): a data-product catalog, docs, and an in-browser query page, so people could find and use the platform.
• Configured ETL pipelines for internal consumers and data products published to the Snowflake Marketplace.
• Wrote internal documentation and engineering updates through Data Mesh Guild posts.

Worked on enterprise data pipelines and analytics infrastructure supporting financial forecasting and modeling workloads.

• Served as technical lead for a team of five engineers from May to July 2022, leading design discussions, code reviews, and implementation planning.
• Built batch data-processing pipelines using Domino, R, Python, Git CI/CD, and Airflow.
• Led the first team in the organization to migrate business-developed models to IT-supported infrastructure, cutting code by 60% and monthly operational effort by 95% (roughly $84K in annual savings per model).
• Developed ETL pipelines using Python, shell scripts, and IBM DataStage.
• Implemented runtime validation and monitoring to maintain data integrity across multiple pipeline stages.
• Orchestrated data workflows across Snowflake, Oracle, and Netezza databases.
• Built visualizations for financial-forecast data using React, D3, Plotly, and Tableau.

• Built data pipelines that moved application data into enterprise data warehouse platforms.
• Implemented validation controls to maintain data integrity across multiple stages of the pipeline.
• Developed ETL jobs using Python, shell scripts, and IBM DataStage.
• Taught quarterly internal training sessions on DataStage and ETL development practices.