Read moreShow lessSelected contributions across Amazon roles+
AI applications & developer platforms
Built a full-stack analytics assistant with a React portal, Python backend, and Redshift/Glue data layer, combining intent routing, RAG, SQL generation, and code execution for charts, reports, and access-controlled answers.
Scaled adoption to 1,000+ stakeholders across 61 teams. Reduced time to insight from 48 hours to 5 minutes for 80% of queries and SQL development time by 40%; enabled no-code onboarding of domains, data sources, and knowledge articles.
Shipped a Redshift Analytics MCP server for schema discovery, query execution, and performance diagnostics, with per-agent identity, audit provenance, and permission guardrails.
Built an autonomous on-call agent for ticket triage, parallel investigations, root-cause analysis, and automatic findings.
Built a SQL code-review agent that executes pull-request queries against Redshift, parses EXPLAIN plans, and surfaces syntax errors and optimization recommendations before merge.
Developed skill and topic authoring, testing, and CI tooling against production-faithful environments; established systematic evaluation of agent output quality.
Created a voice-enabled coaching agent that turns performance data into spoken briefings, plus reusable permission and ticketing frameworks adopted across applications.
Software ownership & data infrastructure
Owned a dispute-reduction workstream from metric definition and Redshift models through pipelines, self-service reporting, customer-facing portal updates, and automated entry validation.
Built a multi-tenant ingestion application across SFTP, web UI, and S3 uploads. Spark validation and Parquet processing supported 100+ feeds across 20 vendors, saving approximately 20 hours per week.
Re-engineered scorecard processing for 3,500 delivery partners and 18 metrics using configurable pipelines, versioning, and Data Vault modeling. Cut runtime from six hours to one and saved approximately 40 hours per week.
Engineered ingestion, analytics, and payment workflows for Alexa Thank My Driver, supporting $2 million in payments; built shared completeness, consistency, and schema checks.
Chewy
Feb 2021 – Mar 2022
Marketing Data Engineer II
Marketing data engineering and machine-learning operations.
+55% customer acquisitions−10% cost per acquisition
Read moreShow lessSelected contributions at Chewy+
Led development of an MLOps platform with SageMaker Studio, custom Docker containers, and Feature Store for model setup, scheduling, monitoring, and updates across AWS accounts.
Built a real-time bid-adjustment solution using competitor pricing for Google Product Listing Ads, increasing customer acquisitions by 55%.
Developed Prophet conversion forecasts that reduced cost per acquisition by 10%.