Johnny Shang
Staff Capacity Engineer @Credit Karma
Signup · Get unlimited contacts
WORK HISTORY
Staff Capacity Engineer @Credit Karma
1) Capacity Forecasting : Use domain knowledge and machine learning models to forecast annual capacity demands for various GCP services and projects, including GCE, GKE, BQ, GCS, etc, and then align with FP & A for budget planning.2) Peak Capacity Sizing : Analyze and estimate seasonal peak capacity demand for each service to prevent capacity issues.3) Cost Efficiency : Identify opportunities for binpacking and rightsizing in GKE to achieve cost savings without impacting service SLA.4) Site Issues : Follow up on site issues and performance troubleshooting to avoid unnecessary capacity additions.5) Performance Optimization : Diagnose and optimize performance issues on services based on Java/Node.js/Typescript.6) Stress Testing : Automate stress testing via Argo Rollouts in prod env without affecting the user experience.7) Capacity Recommendation Engine : Leverage and train machine learning models to implement capacity recommendation engine to streamline capacity rightsizing for cost efficiency.8) Capacity ETL Pipeline : Implement capacity ETL pipeline to collect capacity and performance planning dependent data from multiple sources at appropriate granularity, including K8s performance data, service traffic data, metadata, and more.9) Capacity Knowledge Bot : Implement capacity knowledge bot using LLM on top of collected data by ETL pipeline to simplify and improve communication efficiency across teams.
SKILLS
ABOUT JOHNNY SHANG
Major Valuable Projects in Career ##In Credit Karma, I implemented Capacity Data Pipeline to collect all kinds of data to improve capacity planning efficiency and accuracy which saved lots of engineering efforts. I also accomplished POC (Proof of Concept) to leverage Argo Rollouts in K8s for automated stress testing in prod env and helped dev team on functions design, which reduced stress testing LOE (Level of Effort) from 2 hours to just 1 minute.In Box, I investigated deeply into Intel CPU Spectre/Meltdown vulnerabilities, scheduled canary testing in prod env for different application workloads, and then came up with multiple mitigation approaches which eventually saved Box tens of millions of dollars.To fulfill company level cost saving OKR and accommodate on-prem to public cloud migration, I identified and reclaimed hundreds of zombie servers in different SKU types, I also provided advices to different functional teams to make best use of existing capacity, with all of these we nearly avoided any new server purchase in 2020 and targeting to zero purchase in 2021.In PayPal after split from eBay, I rebuilt the whole capacity team with my manager, expanded team’s coverage from front end capacity planning to IaaS platform, middle tier, database tier and backend storage.In eBay, I built up capacity dedicated data warehouse from scratch, developed Frontend & Database capacity self-service tools to streamline capacity cost/impact estimation.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.