Projects
Selected work
Generalized write-ups of platform work I've led. Details are abstracted for confidentiality — the patterns and outcomes are real.
Multi-account cloud cost governance
FinOps · AWS
Designed and led a cost-optimization initiative spanning close to 100 AWS accounts at a global financial data company. The core architectural decision was using tag-based identity — Application, Environment, Owner, Cost Center — rather than resource IDs, since routine image refreshes constantly change underlying instance IDs and break any tracking built on them.
Built on a hub-and-spoke model with cross-account IAM roles deployed via CloudFormation StackSets, and a parallelized serverless workflow that scans all accounts concurrently and feeds findings into an internal LLM platform for ranked recommendations and executive-ready summaries.
Outcome: a repeatable, self-updating cost governance process that became the primary FinOps reporting mechanism for the platform organization.
Enterprise database consolidation
Migration · SQL Server
Led consolidation of a fragmented SQL Server estate to reduce cloud infrastructure cost and operational overhead, and split a large shared database into isolated clusters to reduce blast radius and improve scalability for independent product teams.
Outcome: materially reduced infrastructure footprint and a cleaner ownership model for each downstream team.
Lakehouse architecture proof of concept
Architecture · Apache Iceberg
Evaluated and prototyped an Apache Iceberg-based lakehouse pattern as a candidate architecture for the next generation of the data platform — assessing it against existing relational workloads for query performance, schema evolution, and cost at scale.
Incident triage acceleration
Reliability · AI-assisted operations
Contributed to an initiative applying AI-assisted triage to production incidents, aimed at reducing mean time to resolution by surfacing likely root causes and relevant runbooks earlier in the incident lifecycle.
Quarterly cloud cost governance program
FinOps · Multi-account audit
Ran a structured, multi-account cost audit combining backup-policy standardization, decommissioning of unused temporary environments, and reserved-capacity right-sizing — the kind of program that only works when every action is documented well enough that a reviewer outside the team can trace each dollar back to a specific decision.
Also caught and flagged a capacity-planning near-miss before a proposed change repeated an earlier costly mistake — preventing, not just fixing, waste.
Outcome: six-figure annual savings in a single quarter, fully auditable, with a governance recommendation adopted for future capacity purchases.
Business-critical event readiness
Reliability · Zero-downtime scaling
Led infrastructure readiness ahead of a major, time-sensitive business event — executing emergency capacity expansions across production and disaster-recovery systems in parallel, with zero service interruption. Authored the incident-response runbook and escalation coverage used during the event window.
Executive-visibility performance investigation
Reliability · Root-cause analysis
Owned the technical investigation into a critical performance issue affecting a top-tier client relationship, under daily leadership review. Isolated the cause to a 24× spike in request volume rather than a system defect — reframing the conversation from "what's broken" to "what changed" and delivering a data-backed resolution path.