Published on 5/14/26

For nearly a decade, the cloud was sold to enterprise CIOs as cheaper, faster, and inherently more flexible than on-premise data centers. The first promise – cheaper – is now widely understood to have been the most overstated. From Wall Street to regional banks, finance executives are confronting cloud bills that have grown faster than their revenue, applications that still take months to release, and architectures that look modern on paper but behave like the legacy systems they were meant to replace.

For Amit Makwana, Senior Application Architect at Softedgelabs LLC, that disconnect has become the defining engineering challenge of his career – and one he has built a measurable track record of solving.

Across recent engagements supporting some of the largest financial institutions in the United States, Makwana’s architectural redesigns have reduced cloud costs by up to 50 percent and accelerated software delivery cycles by two to three times. In an industry where seven-figure cloud overruns and six-month release cadences have quietly become normal, those numbers stand out.

“The cloud isn’t expensive,” Makwana argues. “Cloud architectures that were designed for the old world are expensive. Most of the savings in any modernization program are already sitting inside the system — you just have to find them.”

A Decade of Enterprise Wins

Makwana brings 18 years of enterprise software experience to the cost-and-speed challenge. He has architected mission-critical applications for Federal Home Loan Banks, Bank of America, and Wells Fargo – institutions where downtime is measured in minutes lost, not days, and where every architectural decision must hold up to regulatory scrutiny.

His current focus is Microsoft Azure, the cloud platform of choice for many U.S. financial institutions due to its enterprise integration and compliance posture. In 2024, Microsoft formally certified his Azure capabilities through two credentials: Microsoft Certified: Azure Fundamentals and Microsoft Certified: Azure Developer Associate. The latter validates his ability to design Azure compute solutions, implement Azure security, integrate with third-party services, and -critically for the cost conversation – monitor, troubleshoot, and optimize Azure deployments at scale.

Where the Money Actually Goes

Ask Makwana to explain the 50 percent cost reductions, and he is quick to puncture the assumption that there is any single lever to pull.

“Cloud cost optimization isn’t a discount,” he says. “It’s a redesign.”

In his experience, runaway Azure bills tend to come from a small set of architectural antipatterns: over-provisioned compute, idle workloads running on premium service tiers, chatty microservices generating unnecessary inter-service network traffic, inefficient database queries that scale linearly with load, and caching strategies that were specified but never actually implemented. Each, taken alone, looks minor. Together, they can double a cloud bill within a year of go-live.

His redesigns typically move workloads to the right Azure compute primitive for the actual usage pattern — Azure App Services for sustained traffic, Azure Functions for event-driven spikes, and Azure Kubernetes Service (AKS) only where it genuinely earns its operational complexity. He introduces caching with Redis, tightens database access patterns with Entity Framework Core and SQL Server query tuning, and applies the SOLID principles and CQRS (Command Query Responsibility Segregation), where they reduce coupling and improve testability.

“You don’t need exotic technology to save 50 percent,” he notes. “You need disciplined architecture and the willingness to challenge decisions that nobody has revisited since the initial migration.”

The Second Half of the Equation: Speed

Cost is only half of what Makwana is trying to optimize. The other half – delivery speed – is, in his view, what ultimately determines whether a modernization program is succeeding.

He has been a steady advocate of DevOps maturity inside financial-services engineering organizations, building continuous integration and continuous delivery pipelines on Azure DevOps, GitHub Actions, and Jenkins. He containerizes workloads with Docker, codifies infrastructure with Terraform and ARM templates, and instruments deployments so that engineering leaders can see exactly where in the pipeline their releases are stalling.

The result, in the engagements he describes, is a two-to-three-times reduction in the time it takes for code to move from a developer’s machine to production. That, in turn, gives the business something it has not historically associated with regulated enterprise software: optionality.

“The 2x to 3x delivery acceleration matters because it changes what the business can ask for,” Makwana says. “If a change takes six months, the business stops asking. If it takes six weeks, suddenly there’s a roadmap again.”

A Bigger Conversation: FinOps Meets Architecture

Makwana’s work intersects with one of the most important conversations in U.S. enterprise IT today: how to extract durable value from a decade of cloud migration without throwing more money at the problem.

The discipline now broadly known as FinOps has emerged precisely because organizations realized, often painfully, that cloud bills were not self-regulating. Cost-conscious architecture, observability, and tight feedback loops between finance and engineering have become non-negotiable for large enterprises – particularly in banking, where every basis point of operating expense is scrutinized by analysts and regulators alike.

What sets architects like Makwana apart, peers say, is the combination of technical depth and business literacy. He moves comfortably between system design diagrams and executive conversations, translating architectural choices into the language of cost-to-serve, time-to-market, and risk.

“Most cost overruns aren’t engineering problems – they’re communication problems,” he observes. “Engineers don’t see the bill. Finance doesn’t see the architecture. The architect’s job is to stand in the middle of that conversation and make sure both sides are looking at the same picture.”

Building the Bench

For Makwana, the cost-and-speed agenda is inseparable from the engineering culture that delivers it. He spends significant time mentoring developers, conducting code reviews, and establishing technical standards on the teams he leads – work he considers an essential extension of his architectural role, not a separate activity.

“Architectural diagrams don’t ship code. People do,” he says. “The patterns I draw on a whiteboard are only as good as the engineers who can defend and extend them six months later.”

That focus on building engineering bench strength has practical implications for the financial-services firms he supports. Banks and large enterprises cannot easily hire their way out of a cloud-cost or delivery-speed problem; they must develop the in-house capability to maintain modern architectures responsibly. Increasingly, architects like Makwana are the people responsible for that knowledge transfer.

The Decade Ahead

Looking forward, Makwana sees the architect’s role expanding rather than contracting. As organizations layer AI services, real-time data platforms, and increasingly complex compliance requirements onto already-strained cloud estates, the demand for architects who can hold the whole picture in their heads will grow, not shrink.

“We are about to ask cloud architectures to do a great deal more – agentic services, real-time AI, embedded analytics, all running on top of the same enterprise data,” he says. “The architects who win this decade will be the ones who can introduce that complexity without losing control of cost, security, or delivery speed.”

For now, his focus remains where it has been for nearly two decades: helping U.S. financial institutions ship software that is faster, cheaper, and more resilient than the systems it is replacing — one architectural decision at a time.