As Lead DevOps and Cloud Architect at EPAM Systems on the Amway USA account, I held end-to-end responsibility for the platform’s non-production environments across the USA, Canada, and Dominican Republic (US Region), all environments for Amway Brazil including production, and DevOps processes for the SAP Hybris-based platform. I also led a cross-continental SRE team spanning engineers in the USA, Brazil, the EU, and India. One of headline results: a PoC replacing Akamai with AWS native services cut infrastructure costs by $100,000 per year for the US region alone, improved site access speed by 30%, and was subsequently rolled out across all Amway locations worldwide.
AWS CDN · WAF · Shield
Akamai → AWS migration
Lambda@Edge
SAP Hybris platform
4-continent SRE team
USA · Brazil · EU · India
100k+ redirect rules
Akamai → AWS migration
Lambda@Edge
SAP Hybris platform
4-continent SRE team
USA · Brazil · EU · India
100k+ redirect rules
ClientAmway USA
PlatformSAP Hybris – global retail
RegionsUSA, Canada, Brazil (+ India collaboration)
My roleLead DevOps & Cloud Architect
Key outcome$100k/yr saved · 30% faster · global rollout
$100k
annual cost saving – US region alone
30%
faster site access after CDN switch
100k+
redirect rules migrated via Lambda@Edge
4
continents in the SRE team
Global
Akamai replacement rolled out worldwide
Delivered as Infrastructure Architect at Epam · Epam Certified Technical Mentor · Epam Certified Resource Manager
The Context
A global retail platform with multi-region production complexity
Amway operates one of the world’s largest direct-selling retail networks – a high-availability e-commerce platform serving customers and independent business owners across the USA, Canada, Brazil, and beyond. The SAP Hybris-based platform ran under strict availability, performance, and security requirements. My team owned non-production environments for USA and Canada in full, and held full-stack responsibility – including production — for Amway Brazil. A separate team in India owned USA production, and coordination between the two was a continuous operational requirement.
The SRE team I led spanned four continents
The cross-continental SRE team – L2 and L3 engineers distributed across the USA, Brazil, EU, and India – operated across every time zone simultaneously. This wasn’t a co-located team with remote members. It was a genuinely distributed, multi-cultural team where shift handovers, incident escalation paths, and knowledge transfer all had to work across 12+ hours of time zone difference and multiple languages. Leading this team required building operational processes that worked for people who would never all be in a meeting at the same time.
The Akamai problem: expensive, complex, and hard to evolve
Akamai was the existing CDN and edge security provider – a standard enterprise choice, but expensive and increasingly difficult to evolve alongside the rest of the AWS-based infrastructure. With 100,000+ URL redirection rules embedded in the Akamai configuration, the migration challenge wasn’t just switching providers. Moving that volume of redirect logic to a new platform without breaking the SEO and routing behaviour that Amway’s business depended on was the engineering problem that had kept this migration theoretical. I designed a Lambda@Edge solution that made it practical.
The Akamai → AWS Migration: The Engineering Problem and How It Was Solved
This PoC was the most consequential technical contribution of the engagement – the one that produced measurable, permanent business impact.
1
The redirect problem: 100,000+ rules that couldn’t just be copied
Akamai had accumulated over 100,000 URL redirect rules for Amway USA – the result of years of product catalogue changes, regional routing, campaign landing pages, and SEO management. These rules couldn’t simply be exported and re-imported into CloudFront: AWS CDN doesn’t natively support redirect rule management at that volume with the same pattern-matching flexibility Akamai provides. Without a solution to this specific problem, the migration wasn’t feasible, and this is exactly why the migration had remained a theoretical consideration rather than an executed project.
2
The Lambda@Edge solution: redirect logic executed at CDN edge locations
I designed a Lambda@Edge function that runs redirect logic directly at CloudFront edge locations — before requests ever reach the origin infrastructure. The redirect rules are stored and evaluated at the CDN layer, not the application layer. This solves two problems simultaneously: it makes the 100k+ rule volume manageable in AWS, and it eliminates the origin server load that centralised redirect processing would otherwise create. Requests are redirected at the nearest edge point to the user, which also reduces latency for redirected traffic.
3
The result: 30% faster, $100k saved, and deployed globally
After switching US traffic from Akamai to the AWS native solution (CloudFront CDN + AWS WAF + AWS Shield), access to Amway USA services improved by 30%. The annual infrastructure cost for this part of the stack was reduced by $100,000 – for the US region alone. The PoC validated both the technical approach and the financial case. It was subsequently rolled out across all Amway locations worldwide, multiplying the cost and performance impact across the entire global operation.
The Full Scope of the Engagement
A
Platform DevOps team lead – environment ownership, CI/CD, and operational evolution
Led a four-person Platform DevOps team responsible for all non-production environments for Amway USA and Canada, and all environments including production for Amway Brazil. CI/CD processes for the SAP Hybris-based platform were under our direct management – design, implementation, and ongoing improvement. I worked closely with the India-based production team for Amway USA, coordinating deployments, environment promotions, and architectural changes across both teams.
B
Cross-continental SRE team leadership – L2 and L3 across USA, Brazil, EU, India
Led and mentored the cross-regional SRE team operating L2 (SRE) and L3 (DevOps) support across four continents. Building effective incident response, escalation paths, and knowledge sharing across a team that spans 12+ hours of time zone difference requires explicit operational design – not just good intentions. I built the processes that made the team function reliably across all operating windows, and mentored engineers at both L2 and L3 levels to grow their technical capability and operational maturity.
C
Security incident response – first responder role
Acted as first responder for security incidents across the platform. In a high-volume global retail environment, security events range from attempted credential attacks and bot traffic to WAF rule tuning and DDoS mitigation. Responding effectively requires both security expertise and deep familiarity with the infrastructure – knowing where the risk actually lives and which response actions will contain an incident without disrupting live commerce.
D
Monolith → microservices PoC – validating the architectural direction
In parallel with the Akamai migration, I designed and implemented a PoC for migrating from the SAP Hybris monolith toward a microservices architecture. At a platform of this scale and business criticality, architectural transitions must be validated through working experiments before being committed to at scale. The PoC provided the technical and operational evidence needed for informed architectural decision-making – preventing the organisation from committing to a transformation path without understanding the real-world implementation complexity.
What This Engagement Demonstrates
💰
Cloud cost optimisation with measurable, permanent P&L impact
The Akamai → AWS migration wasn’t framed as a cost-cutting exercise – it was an architectural improvement that happened to reduce costs dramatically. The $100k annual saving for the US region, combined with the 30% performance improvement, is the kind of outcome that gets presented in a quarterly business review. When a PoC produces verified business impact and is rolled out globally, it validates not just the technical approach but the judgement that led to proposing it.
🧩
Solving the specific engineering problem that blocked the migration
100,000+ redirect rules wasn’t a scale problem – it was an architectural problem. AWS CDN doesn’t natively handle that volume and pattern complexity in the way Akamai does. The Lambda@Edge solution I designed was the specific engineering contribution that made the migration possible. Without it, the PoC would have stalled on the same technical barrier that had prevented the migration from being attempted before.
🌍
Leading a genuinely distributed team – not “remote-friendly”, actually global
A cross-continental SRE team where no two members share a time zone is not just a remote team – it’s a different organisational challenge. Incident response, shift handover, escalation paths, and knowledge transfer all need to be explicitly designed to work across 12+ hours of time zone spread. Leading this team produced operational maturity at both L2 and L3 levels across all four regions – not just in the locations I could directly observe.
🔒
Production operations + security incident response on a live commerce platform
A global retail platform is a continuous attack target. Acting as first responder for security incidents while also leading infrastructure modernisation requires the ability to switch context between long-horizon architectural work and immediate operational threat response, and to make the right calls quickly under both sets of conditions. The combination of infrastructure depth and security response capability is less common than either alone.
Skills This Engagement Evidences
Cloud cost optimisation – $100k+ verified savings
Designed and delivered the Akamai → AWS migration PoC producing $100k/year saving (US region) + 30% performance improvement. Subsequently deployed globally. Revenue-visible infrastructure engineering, not theoretical cost modelling.
Lambda@Edge / edge computing architecture
Designed Lambda@Edge solution handling 100k+ redirect rules at CDN edge locations. Eliminated origin server load from redirect processing. The architectural pattern that unblocked the migration entirely.
AWS CDN · WAF · Shield architecture
Replaced Akamai with CloudFront CDN, AWS WAF, and AWS Shield – designed, tested, and validated in production at scale. Deep AWS edge services expertise demonstrated through a production outcome, not a certification.
Cross-continental SRE team leadership
Led L2 + L3 engineers across USA, Brazil, EU, India. Built operational processes for incident response and shift handover that work across 12+ hours of time zone difference. Mentored engineers at both levels across all four regions.
SAP Hybris platform operations
Owned all DevOps processes for a SAP Hybris-based e-commerce platform – non-prod USA/Canada, all environments including production for Brazil. CI/CD design and maintenance. Co-ordination with India production team for USA.
CI/CD design and ownership
Designed and managed CI/CD pipelines across production and lower environments – supporting reliable, repeatable deployments at enterprise scale across multiple regions and teams.
Security incident first response
First responder for security incidents on a global retail platform. WAF rule tuning, DDoS mitigation, credential attack response. Security response combined with infrastructure depth – knowing where the risk actually lives.
Architectural PoC design and validation
Two parallel PoCs: Akamai → AWS (delivered, deployed globally) and monolith → microservices (validated architectural direction). Designing experiments that produce credible evidence rather than confirming assumptions.
The Core Insight from This Engagement
The migrations that don’t happen are often blocked by one specific engineering problem,
not by the migration being infeasible. The Akamai migration had been theoretically desirable
for a long time. The reason it hadn’t happened was 100,000+ redirect rules with no clear path to
AWS. Solving that specific problem – with Lambda@Edge – unlocked the entire migration and produced
$100k/year in savings as a direct consequence.
The second insight: distributed team performance is a design problem, not a culture problem.
A cross-continental SRE team doesn’t fail because people in different time zones don’t communicate well.
It fails because the operational processes – incident escalation, shift handover, runbook access,
on-call thresholds – weren’t designed for a distributed context. Getting those right is an
engineering and management discipline, not a motivation question.
The third: in a PoC, the goal is not to prove that an idea works in a test environment.
It is to surface the specific constraints that will prevent it from working at scale –
before those constraints become production incidents.
Who This Is Relevant For
You’re paying for Akamai and wondering whether AWS native services could replace it
The question isn’t whether AWS CDN, WAF, and Shield are technically capable – they demonstrably are. The question is how to handle the migration complexity at scale: the redirect rules, the configuration parity, the traffic cutover without disruption. This engagement shows what that migration looks like in a production global retail environment with $100k+ annual cost impact.
You have a CDN with thousands of redirect rules that feel impossible to migrate
Large redirect rule sets are one of the most common practical blockers in CDN migrations. The Lambda@Edge pattern – handling redirect logic at edge locations rather than at the origin – is a specific architectural solution to this specific problem. If you have a similar situation, the design is documented and the production result is verified.
You’re leading or building a distributed DevOps or SRE team across multiple regions
Cross-regional teams work when the processes are designed correctly – shift handover, escalation paths, on-call coverage, runbook access. They struggle when these are designed for a co-located team and then applied to a distributed one. The team structure here – USA, Brazil, EU, India – required explicit operational design across every time zone.
You need infrastructure modernisation without disrupting live commerce
Every change on a live e-commerce platform carries business risk. The approach here – validating architectural changes through production-grade PoCs before committing to them at scale – is the methodology that makes large-platform modernisation predictable rather than dangerous.
Your engineering leadership needs to show P&L impact, not just delivery metrics
Infrastructure investments that produce verified, quantified business outcomes – cost reduction, performance improvement, or both – are the ones that build engineering credibility with business leadership. The $100k/year result here was directly presented in business terms, not buried in infrastructure metrics.
You’re on a SAP Hybris platform and need to evolve it without starting over
SAP Hybris platforms carry significant business logic, integration complexity, and organisational investment. The approach here – using PoCs to explore microservices decomposition in parallel with operational improvements – shows how to move a Hybris-based platform forward without a big-bang rewrite.
Optimising a CDN or edge infrastructure spend, building distributed SRE teams, or modernising a large retail or e-commerce platform?
In 30 minutes I can tell you whether your situation maps to what I do – the real costs in your current CDN setup, the specific engineering problems blocking your migration, or the process gaps in your distributed team.
No pitch. A direct conversation about your infrastructure, your team, and where the real leverage is.