NETGEAR SUBSCRIPTION SYSTEM
Principal Engineer & Cloud Architecture Lead (2022–2023), reporting to the CTO
Principal Engineer & Cloud Architecture Lead (2022–2023), reporting to the CTO
Modernizing the subscription platform behind millions of customers and tens of millions of devices, with no big-bang cutover.
THE CHALLENGE :
THE CHALLENGE :
THE CHALLENGE :
NETGEAR's subscription systems had grown organically under Conway's Law. The result was four conflicting systems of record, fragmented data, and weak governance.
Engineering, IT, customer service, finance, and marketing all depended on that data, and so did the website, the mobile app, routers worldwide, and the Armor and Smart Parental Controls security products. A full rewrite with a cutover weekend was too risky. The job was to put clean architecture and governance in place without disrupting the business.
APPROACH & KEY DECISIONS :
APPROACH & KEY DECISIONS :
APPROACH & KEY DECISIONS :
A green-field rebuild was technically straightforward but organizationally dangerous. I focused on reducing risk while still delivering a fundamentally better system.
Key decisions included:
Strangler pattern: put the new architecture behind the existing APIs so it could replace the old systems piece by piece
Shadow mode: ran the new system alongside the old one and compared results before it took over
Sequenced the takeover: new subscriptions first, then renewals, then the remaining data, so there was never a big-bang migration
Created reusable standards for testing, pipelines, and infrastructure as code that other teams could adopt
Invested heavily in aligning stakeholders who had competing priorities
SOLUTION & IMPLEMENTATION :
SOLUTION & IMPLEMENTATION :
SOLUTION & IMPLEMENTATION :
I led the design and delivery across engineering, IT, and business teams, directly leading senior engineers alongside four delivery teams (20+ engineers, four Engineering Managers, and a Senior Director).
A governed subscription service became the single source of truth for the company
Consumer teams kept their familiar APIs while the backend was replaced underneath them
New CI/CD, automated testing, and Terraform-based infrastructure improved reliability and speed
API-first design (Swagger) and a joint Product–Engineering intake shortened the path from idea to build
Clear communication and risk management brought initially resistant stakeholders along
RESULTS & IMPACT :
RESULTS & IMPACT :
RESULTS & IMPACT :
Service-delivery accuracy raised from ~90% to over 99%
API response times cut from over 1 second to under 100 ms
Infrastructure cost cut 22% through AWS modernization, Terraform CI/CD, and Kubernetes, and a further 31% on the peak-event fleet through auto-scaling
Product-to-Engineering handoff cut from 2–3 weeks to a week or less
One source of truth for subscription data across the company, delivered with minimal disruption
Cloud and backend standards adopted company-wide across a 140+ engineer organization
Mentored three Engineering Managers and more than a dozen engineers who were promoted
WHAT I LEARNED :
WHAT I LEARNED :
WHAT I LEARNED :
Large-scale platform change is rarely just a technical problem. Building shared understanding and trust across previously siloed teams was harder than the engineering. Technical elegance only succeeds when it's paired with deliberate stakeholder management and risk reduction.
TECHNOLOGIES :
TECHNOLOGIES :
TECHNOLOGIES :
AWS · Microservices · Strangler pattern · Data governance · Infrastructure-as-Code (Terraform) · Kubernetes · CI/CD · API-first design
NETGEAR SUBSCRIPTION SYSTEM
Principal Engineer & Cloud Architecture Lead (2022–2023), reporting to the CTO
Principal Engineer & Cloud Architecture Lead (2022–2023), reporting to the CTO
Modernizing the subscription platform behind millions of customers and tens of millions of devices, with no big-bang cutover.
THE CHALLENGE :
THE CHALLENGE :
THE CHALLENGE :
NETGEAR's subscription systems had grown organically under Conway's Law. The result was four conflicting systems of record, fragmented data, and weak governance.
Engineering, IT, customer service, finance, and marketing all depended on that data, and so did the website, the mobile app, routers worldwide, and the Armor and Smart Parental Controls security products. A full rewrite with a cutover weekend was too risky. The job was to put clean architecture and governance in place without disrupting the business.
APPROACH & KEY DECISIONS :
APPROACH & KEY DECISIONS :
APPROACH & KEY DECISIONS :
A green-field rebuild was technically straightforward but organizationally dangerous. I focused on reducing risk while still delivering a fundamentally better system.
Key decisions included:
Strangler pattern: put the new architecture behind the existing APIs so it could replace the old systems piece by piece
Shadow mode: ran the new system alongside the old one and compared results before it took over
Sequenced the takeover: new subscriptions first, then renewals, then the remaining data, so there was never a big-bang migration
Created reusable standards for testing, pipelines, and infrastructure as code that other teams could adopt
Invested heavily in aligning stakeholders who had competing priorities
SOLUTION & IMPLEMENTATION :
SOLUTION & IMPLEMENTATION :
SOLUTION & IMPLEMENTATION :
I led the design and delivery across engineering, IT, and business teams, directly leading senior engineers alongside four delivery teams (20+ engineers, four Engineering Managers, and a Senior Director).
A governed subscription service became the single source of truth for the company
Consumer teams kept their familiar APIs while the backend was replaced underneath them
New CI/CD, automated testing, and Terraform-based infrastructure improved reliability and speed
API-first design (Swagger) and a joint Product–Engineering intake shortened the path from idea to build
Clear communication and risk management brought initially resistant stakeholders along
RESULTS & IMPACT :
RESULTS & IMPACT :
RESULTS & IMPACT :
Service-delivery accuracy raised from ~90% to over 99%
API response times cut from over 1 second to under 100 ms
Infrastructure cost cut 22% through AWS modernization, Terraform CI/CD, and Kubernetes, and a further 31% on the peak-event fleet through auto-scaling
Product-to-Engineering handoff cut from 2–3 weeks to a week or less
One source of truth for subscription data across the company, delivered with minimal disruption
Cloud and backend standards adopted company-wide across a 140+ engineer organization
Mentored three Engineering Managers and more than a dozen engineers who were promoted
WHAT I LEARNED :
WHAT I LEARNED :
WHAT I LEARNED :
Large-scale platform change is rarely just a technical problem. Building shared understanding and trust across previously siloed teams was harder than the engineering. Technical elegance only succeeds when it's paired with deliberate stakeholder management and risk reduction.
TECHNOLOGIES :
TECHNOLOGIES :
TECHNOLOGIES :
AWS · Microservices · Strangler pattern · Data governance · Infrastructure-as-Code (Terraform) · Kubernetes · CI/CD · API-first design
NETGEAR SUBSCRIPTION SYSTEM
Principal Engineer & Cloud Architecture Lead (2022–2023), reporting to the CTO
Principal Engineer & Cloud Architecture Lead (2022–2023), reporting to the CTO
Modernizing the subscription platform behind millions of customers and tens of millions of devices, with no big-bang cutover.
THE CHALLENGE :
THE CHALLENGE :
THE CHALLENGE :
NETGEAR's subscription systems had grown organically under Conway's Law. The result was four conflicting systems of record, fragmented data, and weak governance.
Engineering, IT, customer service, finance, and marketing all depended on that data, and so did the website, the mobile app, routers worldwide, and the Armor and Smart Parental Controls security products. A full rewrite with a cutover weekend was too risky. The job was to put clean architecture and governance in place without disrupting the business.
APPROACH & KEY DECISIONS :
APPROACH & KEY DECISIONS :
APPROACH & KEY DECISIONS :
A green-field rebuild was technically straightforward but organizationally dangerous. I focused on reducing risk while still delivering a fundamentally better system.
Key decisions included:
Strangler pattern: put the new architecture behind the existing APIs so it could replace the old systems piece by piece
Shadow mode: ran the new system alongside the old one and compared results before it took over
Sequenced the takeover: new subscriptions first, then renewals, then the remaining data, so there was never a big-bang migration
Created reusable standards for testing, pipelines, and infrastructure as code that other teams could adopt
Invested heavily in aligning stakeholders who had competing priorities
SOLUTION & IMPLEMENTATION :
SOLUTION & IMPLEMENTATION :
SOLUTION & IMPLEMENTATION :
I led the design and delivery across engineering, IT, and business teams, directly leading senior engineers alongside four delivery teams (20+ engineers, four Engineering Managers, and a Senior Director).
A governed subscription service became the single source of truth for the company
Consumer teams kept their familiar APIs while the backend was replaced underneath them
New CI/CD, automated testing, and Terraform-based infrastructure improved reliability and speed
API-first design (Swagger) and a joint Product–Engineering intake shortened the path from idea to build
Clear communication and risk management brought initially resistant stakeholders along
RESULTS & IMPACT :
RESULTS & IMPACT :
RESULTS & IMPACT :
Service-delivery accuracy raised from ~90% to over 99%
API response times cut from over 1 second to under 100 ms
Infrastructure cost cut 22% through AWS modernization, Terraform CI/CD, and Kubernetes, and a further 31% on the peak-event fleet through auto-scaling
Product-to-Engineering handoff cut from 2–3 weeks to a week or less
One source of truth for subscription data across the company, delivered with minimal disruption
Cloud and backend standards adopted company-wide across a 140+ engineer organization
Mentored three Engineering Managers and more than a dozen engineers who were promoted
WHAT I LEARNED :
WHAT I LEARNED :
WHAT I LEARNED :
Large-scale platform change is rarely just a technical problem. Building shared understanding and trust across previously siloed teams was harder than the engineering. Technical elegance only succeeds when it's paired with deliberate stakeholder management and risk reduction.
TECHNOLOGIES :
TECHNOLOGIES :
TECHNOLOGIES :
AWS · Microservices · Strangler pattern · Data governance · Infrastructure-as-Code (Terraform) · Kubernetes · CI/CD · API-first design