Shaun Thomas Email & Phone Number

Normal, IL, US

Cincinnati, Ohio, US

• Overhauled client's OLAP ETL procedure using strategic partitioning, columnar storage, unlogged tables, and incremental processing to reduce processing time from 12 hours to 1 hour.
• Audited client's 100k-line PL/pgSQL stored procedure business logic and application queries to eliminate performance bottlenecks by adding missing indexes, converting procedures to CTEs, and optimizing query execution.
• Initiated "Tembo University" project to drive developer interest through demonstration of sophisticated Postgres techniques including locking, queues, asynchronous calls, vectors, semantic search, and geographical.
• Integrated automated CloudNativePG declarative hibernation functionality into the Tembo Kubernetes Rust operator to reduce cloud provider billing for idle instances by 90%.
• PostgreSQL advocacy through regular blogging and community interaction.

Bedford, MA, US

• Designed 3-node synchronous cross-region architecture using PgBench and EFM on AWS for a financial client and demonstrated remote-apply mode only added 15% commit-time overhead at 150 sessions to maximize VPS CPU.
• Simplified the software stack for a financial client by replacing Redis with a Postgres cache table updated through PL/pgSQL trigger logic, increasing total TPS by 2.5x and reducing execution latency from 7.3 to 3.3ms.
• Led engineering team for product rewrite of the High Availability Redesign for Postgres (HARP) quorum-driven connection routing system for Enterprise Postgres Distributed (PGD) to reduce Multi-Master write conflicts.
• Introduced six Postgres HA architectures as new standard for all company documentation, deployments, and sales engagements for consistent messaging for HA best practices.
• Demonstrated through automated benchmarking and analysis that increasing max_wal_size can provide up to 3x better TPS throughput by reducing storage contention and improving write spreading.
• Designed and delivered 3-part internal training course on Patroni, covering installation, dependencies such as etcd and HAProxy, operation, and troubleshooting, to T1-T3 support engineers for assisting customers with.

Oxford, Oxfordshire, GB

• Co-authored "Always-On" Multi-Master Postgres cluster architecture optimized for immediate switchover for lowest demonstrable RTO of 5 seconds (five 9s), online major-version upgrades, reduced divergent conflict risks.
• Innovated new Postgres catalog manipulation technique using PL/pgSQL functions and pg_cron to monitor and concurrently remove index bloat on 20TB client cluster to eliminate offline maintenance.
• Led a client migration from a 3-node repmgr cluster to a 4-node Always-On Multi-Master BDR architecture, including extensive testing and validation.
• Developed a 3-node Postgres repmgr cluster for a client in a high-concurrency microservice environment specifically to accommodate frequent VM recycling for compliance requirements.
• Designed and deployed a 3-node Postgres repmgr cluster with Barman integration, including necessary support scripts and playbooks, suitable for a client with low on-site Postgres expertise.
• Audited a client's Postgres infrastructure to identify and eliminate configuration, role, privilege, SQL, and connection vulnerabilities to harden it against malicious users or unknown exploits.

Chicago, IL, US

• Innovated Django-driven Postgres cluster orchestration tool for Linux servers named Elephant Herd as a Service to manage HA, monitoring, rebuild, and backups for over 100 server instances across financial data hubs in.
• Engineered pure Linux Bash backup system leveraging hard links and parallel compression (pigz) to produce full and incremental backups of Postgres clusters, reducing backup time of 50TB system from 8-12 hours to 30.
• Deployed partition-driven mechanism using Postgres FDW to migrate expired data to long-term 50TB VLDB OLAP archival and reporting cluster to preserve efficiency of source instances.
• Produced and delivered 3-day training course for SQL Server DBAs at Apex Clearing on Postgres internals and features so they could safely interact with the Peak6 Postgres infrastructure.
• Engineered Saltstack pillar to automate Postgres infrastructure deployment by VM designation and simplify cluster management.
• Derived Postgres hardware infrastructure roadmap based on 3-year storage growth projections and CPU / memory saturation statistics for over 20 clusters to ensure long-term viability.

Chicago, IL, US

• Established High-Availability Postgres cluster using Linux, DRBD, LVM, and Pacemaker for 99.99% uptime of a 35k TPS / 1-billion transaction-per-day options trading platform at Peak6. Premiered the stack as a live.
• Replaced 10-spindle HDD RAID-10 with FusionIO NVMe PCIe storage to improve cold system reads by over 90% and eliminate application timeouts following Postgres restarts on options trading platform. Presented as an.
• Standardized organization SOPs for database-related topics such as: emergency checklists and playbooks, instance architecture, and disaster recovery, to ensure accurate representation and knowledge transfer.
• Launched internal education initiative to familiarize developers with Postgres best practices, system internals, edge cases, and innovative techniques, ensuring better database interoperability.
• Implemented daily Postgres backup instance recovery and sensitive data redaction in staging and dev environments for backup validation and to provide production-scale testing capabilities.
• Led hiring, training, and management for Postgres DBA team for reserve capacity and coverage purposes.

Chicago, IL, US

• Managed and optimized Postgres and MS SQL Server databases for high-volume ad impression statistics, aggregation, and reporting.
• Developed ETL processes in Python, Perl, and Bash for data synchronization and reporting.
• Designed star/snowflake schema OLAP Postgres system for critical internal analytics and customer reports.
• Established all internal Postgres database SOPs and documentation for system maintenance, upgrades, monitoring, and architecture, to ensure process accountability.
• Developed Nagios abstraction library for Python so developers could integrate data statistics, monitoring, and event logging directly in the application layer.
• Collaborated with infrastructure team to consolidate and decommission unused equipment for a cost savings in excess of $100k YoY.

Moline, IL, US

• Administered MySQL, Postgres, and Sybase databases, optimizing for performance and scalability.
• Developed an AI-driven Bayesian digital archive categorization system using PHP for long-term PDF newspaper distribution.
• Built a content-parsing engine in PHP for extracting metadata from text automotive classified ads and several proprietary vendor formats for automotive inventory system.
• Produced templated database-driven retail website engine in PHP for over 100 newspapers and independent lots to publish automotive inventory.
• Standardized company source control on CVS for better team code collaboration.