AI-Powered Modernization at AWS: Leading the Evolution from Machine Learning to Agentic Systems
A multi-year transformation toward intelligent, trustworthy modernization experiences
A multi-year transformation toward intelligent, trustworthy modernization experiences
As an AWS employee, I adhere to company policies regarding data confidentiality. While I cannot share specific usage metrics or proprietary customer adoption data, I have provided qualitative insights and outcomes to demonstrate the impact of my contributions. These examples align with publicly available information and best practices while respecting internal guidelines.
Modernizing .NET applications is one of the most complex transformations enterprise customers undertake. When I inherited this domain, the experience was fragmented across disconnected tools, inconsistent mental models, and limited guidance. Customers struggled to move from initial assessment to code transformation with confidence.
I led the UX strategy across four major modernization products that together represent the evolution of AI at AWS:
1. AWS Transform for .NET (Agentic AI)
2. Microservice Extractor (AI-assisted workflows)
3. Migration Hub Strategic Recommendations (Machine learning insights)
4. Modernization Calculator (Unified modernization cost model)
Each tool built on the one before it. Together, they created the first cohesive modernization journey at AWS and laid the foundation for the agentic future.
AWS Transform (formerly Q Code Transformation) enables enterprises to port .NET applications from Windows to Linux up to four times faster, reducing operational costs by as much as 40 percent. Transform analyzes code, maps dependencies, and proposes refactoring pathways with minimal human intervention while preserving customer oversight.
I took ownership of Transform four months before its planned re:Invent 2024 launch. At that point, the project had no approved requirements, minimal UX support, and no pipeline for staffing due to reorg shifts and hiring freezes. To move forward, I redeployed a designer from my AWS Messaging team, created a structured onboarding plan for this high-ambiguity space, and established a predictable working cadence across engineering and product.
My leadership contributions included the following:
I aligned .NET, VMware, and Mainframe teams around a unified agentic modernization experience.
I conducted accelerated CX Bar Raising Reviews to surface and eliminate adoption risks early.
I defined trust and transparency mechanisms that helped customers understand how and when the agent rewrites code.
I shaped workflows that balanced automated orchestration with human oversight.
I ensured that mental models established in Microservice Extractor and AWS Toolkit carried forward seamlessly into the agentic experience.
I coached the team to shift from a traditional tool mindset to an agentic thinking model.
Despite accelerated timelines, AWS Transform launched successfully and was featured in the AWS Chief Executive Officer keynote at re:Invent 2024. It became AWS’s first agentic experience and set the foundation for future AI-driven workflows across the organization.
Microservice Extractor enables cloud architects and developers to decompose large monolithic applications into microservices. It scales to applications with up to 50,000 classes and integrates with the AWS Toolkit so that developers can move extracted components directly into service development.
Cloud architects and developers approach modernization very differently. Architects think about system boundaries, domain-driven groupings, dependency separation, and overall application structure. Developers think about file-level refactoring, method dependencies, code correctness, and efficient execution. My role as the UX leader was to bridge these perspectives and ensure the experience supported both mental models in a unified way.
I provided strategic direction in the following ways:
I defined the visualization strategy for representing large dependency graphs so that architects could interpret system structures accurately.
I guided designers to create grouping and regrouping interactions that preserved architectural intent while giving developers flexibility to refine details.
I introduced transparency mechanisms that clarified the accuracy and limitations of machine learning generated groupings.
I coordinated the workflow between Microservice Extractor and the .NET Toolkit to connect architectural decomposition with practical code transformation.
I ensured that the mental models established in this tool informed the orchestration patterns later used in AWS Transform.
Microservice Extractor became the bridge between modernization planning and code execution. It created the groundwork that enabled agentic orchestration in AWS Transform.
Migration Hub Strategic Recommendations (MHSR) helps enterprises analyze large set of applications and identify modernization paths.
I led the product through three major phases of its evolution:
Phase 1: Establishing the foundation
I defined the design strategy for the initial release, simplifying complex infrastructure assessments into clear recommendations. This gave enterprises their first unified view of their application portfolios.
Phase 2: Expanding through source code analysis
As the product expanded to include source code and database analysis, I partnered closely with UX research to understand customer concerns about sensitive data. I helped design transparency mechanisms that increased trust and clarified data workflows.
Phase 3: Introducing machine learning generated groupings
For customers managing thousands of interconnected applications, we introduced machine learning generated groupings to help them understand modernization pathways. I defined the strategy for visualizing these groupings in a way that balanced cost, technical complexity, and organizational structure.
MHSR established the first intelligence layer in the modernization portfolio, enabling customers to navigate modernization with clarity and confidence.
The Modernization Calculator was the first AWS calculator designed around customer mental models rather than infrastructure attributes. Traditional calculators focus on instance types or storage classes, but modernization customers think about cost through the lens of their application architecture, workload patterns, and modernization paths.
As the UX leader, I guided the experience strategy that shifted AWS toward a modernization-focused cost model. This required coordinating multiple AWS service teams to unify their pricing inputs into a coherent workflow.
I contributed in the following ways:
I established the conceptual framework that brought multiple AWS services together into a single modernization cost model.
I ensured the workflow matched how architects evaluate modernization decisions instead of mirroring internal AWS pricing structures.
The Modernization Calculator became the entry point into the modernization journey and shifted how AWS teams positioned modernization offerings to customers.
Throughout this program, I applied leadership mechanisms that elevated customer experience and improved product quality.
Led CX Bar Raising Reviews across all modernization tools to ensure alignment with customer mental models and experience standards.
Created education programs that helped PMs and engineers understand how customer experience directly influences modernization adoption.
Secured dedicated UX research support in a space where researchers had previously struggled to gain traction.
Guided foundational studies, including a re:Invent 2024 initiative focused on customer trust in AI systems.
Evaluated tradeoffs between immediate customer benefit and long-term UX debt, and created mechanisms to ensure that debt was revisited and addressed.
Navigated organizational challenges such as layoffs, hiring freezes, and staffing gaps while maintaining delivery momentum and protecting teams from burnout.
Facilitated knowledge sharing across modernization efforts to ensure alignment and raise the experience bar across all products.
Accelerated modernization timelines for enterprise customers by up to four times.
Reduced operational costs by up to 40 percent in Windows-to-Linux modernization scenarios.
Created experience framework that scale for customers managing thousands of applications.
Established transparency and oversight models that strengthened trust in AI driven and agentic workflows.
Created organizational mechanisms, including CX office hours and Bar Raising Reviews, that scaled across AWS.
Delivered AWS Transform, which was featured in the AWS Chief Executive Officer keynote at re:Invent 2024.
The journey from MHSR to Microservice Extractor to AWS Transform mirrors the evolution of AI capabilities at AWS, moving from machine learning insights to AI assisted workflows and finally to fully agentic systems. My role was to ensure that each step of this evolution reflected customer mental models, responsible AI design, and scalable mechanisms teams could rely on.
This work strengthened my belief that trustworthy AI experiences require clear intent, transparent systems, and mechanisms that balance automation with thoughtful human oversight. The modernization portfolio reflects some of the most meaningful opportunities for AI to transform enterprise software, and I am proud to have helped shape that direction.
Fig 1: QCT .NET: Create transformation job
Fig 2: QCT .NET: Create transformation job
Fig 3: QCT .NET: Create transformation job
Fig 4: QCT .NET: Transformation job in progress
Fig 5: QCT .NET: Selecting repos for transformation
Fig 6: QCT .NET: Send repo selection for approval
Fig 7: QCT .NET: Approve repo selection for transformation
Fig 8: QCT .NET: Transformation Dashboard
Fig 9: QCT .NET: Transform in Visual Studio
Fig 10: QCT .NET: Transform in Visual Studio
Fig 11: Microservice Extractor - Visualization canvas of class grouping
Fig 12: .NET Modernization Calculator - Select architecture
Fig 13: .NET Modernization Calculator - Select architecture size
Fig 14: .NET Modernization Calculator - Select architecture patterns
Fig 15: .NET Modernization Calculator - Edit configuration
Fig 16: .NET Modernization Calculator -Estimates
Fig 17: Strategy Recommendation - EC2 assessment