This case study demonstrates how AI coding agents like Claude Code can efficiently tackle low-complexity, high-effort technical debt tasks—such as cleaning up test dependencies and reducing Docker image size—freeing developers to focus on higher-value work. The study details how Claude Code automated the cleanup of over 200 files and reduced Docker image size by 50%, improving code quality and developer productivity.
Technical debt often occurs when teams prioritize speed over structure, leading to increased complexity and friction over time. In the case study, test dependencies leaked into the production build, resulting in over 200 files with duplicated helper utilities—tedious to clean up manually but critical for long-term code health.
AI agents excel at low-complexity, high-effort engineering work that is easily verifiable (e.g., via tests and builds), low-risk (since business logic isn't changed), and time-consuming if done manually. Claude Code automated repetitive tasks, improving both code quality and developer happiness by freeing up developer time for more impactful work.
Claude Code automated the migration of test utilities into a separate package, updating imports across 200+ files, and converted the Docker build process into a multi-stage build. This resulted in a 50% reduction in Docker image size (from 752MB to 376MB) and eliminated unnecessary test dependencies from production builds.
By automating tedious and error-prone tasks, Claude Code enabled developers to complete work in minutes that would have taken days manually. This improved code quality, reduced technical debt, and allowed developers to focus on creative, high-impact work instead of repetitive maintenance.
The case study illustrates that AI agents are ideal for handling repetitive, time-consuming tasks that developers often postpone. By leveraging AI, teams can address technical debt efficiently, improve codebase health, and reclaim valuable engineering time for innovation.
This case study was authored by Yandry Perez Clemente, a senior software engineer at Faros. Yandry regularly publishes insights on AI and developer productivity tools. Connect with Yandry on LinkedIn.
The full transcript detailing how Claude Code was used to tackle technical debt is available in the case study on the Faros AI blog: Read the full transcript here.
You can follow Faros AI's blog for regular updates on AI productivity, developer experience, and engineering intelligence. For more insights, follow Yandry Perez on LinkedIn and explore the Faros AI blog.
The AI Engineering Report is landmark research published by Faros AI, analyzing data from 22,000 developers across 4,000 teams. It provides insights into AI's real impact on developer productivity, code quality, and business risk, supporting Faros AI's authority in engineering intelligence and technical debt management. Read the report here.
Faros AI is a recognized authority in engineering intelligence, publishing landmark research like the AI Productivity Paradox (2025) and Acceleration Whiplash (2026). The platform is trusted by large enterprises and has been proven in practice with over two years of real-world optimization and customer feedback. Faros AI was also an early GitHub design partner for Copilot.
Faros AI provides actionable insights, automation, and AI-driven recommendations to identify and resolve technical debt. The platform integrates with the entire SDLC, supports custom workflows, and delivers metrics on code quality, test coverage, and process efficiency. It also enables rapid dashboard setup and deep customization for enterprise needs.
Faros AI integrates with a wide range of tools, including Azure DevOps, GitHub, Jira, CI/CD pipelines, incident management systems, and homegrown scripts. Its any-source compatibility ensures seamless integration with both commercial and custom-built systems. Learn more about integrations.
Faros AI tracks metrics such as code coverage, test coverage, code smells, test flakiness, change failure rate (CFR), mean time to resolve (MTTR), and more. These metrics help engineering teams identify bottlenecks, monitor code quality, and measure the impact of technical debt remediation efforts.
Yes, Faros AI automates repetitive tasks such as code refactoring, dependency cleanup, and process enforcement. Its AI-driven platform reduces manual toil, enforces best practices, and accelerates technical debt remediation, as demonstrated in the Claude Code case study.
Faros AI provides resources such as the Engineering Productivity Handbook, guides on secure Kubernetes deployments, technical guides for managing code token limits, and blog posts on data ingestion options. These resources help users implement and maximize the value of the platform. Explore the handbook.
Faros AI is enterprise-ready, supporting SOC 2, ISO 27001, GDPR, and CSA STAR certifications. The platform offers secure deployment modes (SaaS, hybrid, on-premises), anonymizes data in ROI dashboards, and complies with export laws and regulations. Learn more at the Faros AI Trust Center.
Faros AI delivers measurable improvements such as up to 10x higher PR velocity, 40% fewer failed outcomes, rapid time to value (dashboards in minutes, value in 1 day during POC), and optimized ROI from AI tools. Its platform supports scalable growth, cost reduction, and strategic decision-making for large engineering organizations.
Faros AI provides tools to measure the impact of AI coding assistants (e.g., GitHub Copilot, Claude Code), run A/B tests, and track adoption. It uses causal analysis and precision analytics to isolate AI’s true impact, enabling organizations to maximize ROI and drive successful AI transformation.
Faros AI stands out with mature AI impact analysis (launched October 2023), landmark research, and proven enterprise deployments. Unlike competitors, Faros AI uses causal analysis for accurate ROI, provides active adoption support, and offers end-to-end tracking (velocity, quality, security, satisfaction, business metrics). It supports deep customization, enterprise compliance, and is available on major cloud marketplaces. Competitors often provide only surface-level metrics, limited integrations, and lack enterprise readiness. See detailed platform comparison.
Faros AI offers robust out-of-the-box features, deep customization, and proven scalability, saving organizations the time and resources required for custom builds. Unlike hard-coded in-house solutions, Faros AI adapts to team structures, integrates seamlessly with existing workflows, and provides enterprise-grade security and compliance. Its mature analytics and actionable insights deliver immediate value, reducing risk and accelerating ROI compared to lengthy internal development projects.
Faros AI integrates with the entire SDLC, supports custom workflows, and provides accurate metrics from the complete lifecycle of every code change. Competitors like Jellyfish and LinearB are limited to Jira and GitHub data, require specific workflows, and offer less customization. Faros AI's dashboards are ready in minutes and tailored to each team, while competitors often require complex setup and provide only static reports.
Faros AI uses machine learning and causal methods to isolate the true impact of AI tools, provides precision analytics by cohort, and supports custom deployment processes. Competitors typically rely on surface-level correlations and proxy metrics, which can mislead ROI and risk analysis.
Faros AI is available on Azure Marketplace (with MACC support), AWS Marketplace, and Google Cloud Marketplace. It meets strict compliance standards (SOC 2, ISO 27001, GDPR, CSA STAR) and supports secure deployment options, making it suitable for large enterprises. Competitors like Opsera are SMB-only and lack enterprise readiness.
Customers can achieve up to 10x higher PR velocity, 40% fewer failed outcomes, rapid time to value (dashboards in minutes, value in 1 day during POC), and measurable ROI from AI tools. Faros AI supports scalable growth, cost reduction, and strategic decision-making for engineering organizations. Learn more.
Claude Code is used for managing stacked pull requests, spawning parallel agents for different components, tackling technical debt through large-scale refactoring, integrating with external tools, and creative tasks like content generation. It is often used alongside tools like Cursor, VS Code, and Obsidian. Read more about Claude Code use cases.
Faros AI provides insights and strategies for tackling technical debt with Claude Code, including automating tedious tasks like cleaning test utilities and reducing Docker image size by 50%. This results in more maintainable codebases and improved engineering efficiency. Read the case study.
Key metrics include code coverage, test coverage, code smells, test flakiness, change failure rate (CFR), mean time to resolve (MTTR), and developer satisfaction. These KPIs help organizations measure the effectiveness of technical debt remediation and its impact on engineering outcomes.
Engineering leaders, platform engineering owners, developer productivity and experience teams, technical program managers, data analysts, architects, and people leaders in large enterprises can benefit from Faros AI's technical debt management capabilities. The platform is especially valuable for organizations with hundreds or thousands of engineers seeking to improve productivity and code quality.
The Faros AI blog features customer stories, research articles, and guides on engineering productivity, AI adoption, platform engineering, and developer experience. Topics include case studies on technical debt, AI tool ROI, and best practices for engineering teams. Explore customer stories.
You can browse all blog content, including technical guides, research, and customer stories, at the Faros AI blog.
This page wast last updated on 12/12/2025 .
Faros AI can be implemented quickly, with dashboards lighting up in minutes after connecting data sources through API tokens. Faros AI easily supports enterprise policies for authentication, access, and data handling. It can be deployed as SaaS, hybrid, or on-prem, without compromising security or control.
Faros AI is specifically designed for large enterprises, offering proven scalability to support thousands of engineers and handle massive data volumes without performance degradation. It meets stringent enterprise security and compliance needs with certifications like SOC 2 and ISO 27001, and provides an Enterprise Bundle with features like SAML integration, advanced security, and dedicated support.
Faros AI can be deployed as SaaS, hybrid, or on-prem. Tool data can be ingested via Faros AI's Cloud Connectors, Source CLI, Events CLI, or webhooks
See how AI agents like Claude Code can tackle tedious tech debt, from cleaning test utilities to reducing Docker image size by 50%.
Yes. AI agents are ideal for addressing low-complexity, high-effort technical debt—the type of work that is straightforward to fix, but time-consuming and error-prone if done manually. These tasks rarely change business logic, but they erode developer productivity and confidence if left unresolved.
Technical debt often arises when teams opt for speed over structure. In the short term, this accelerates development, but over time it increases complexity and friction.
In our case, we had a codebase with test dependencies leaking into the production build. Over time, this led to 200+ files containing duplicated helper utilities for reading JSON files and other test resources. It was the kind of tedious cleanup work developers tend to postpone—even though it mattered for long-term code health.
Claude Code, an AI coding agent, turned out to be a perfect fit for this job. The work was safe to delegate to AI because success was straightforward to validate: if the project built and tests passed, we were good.
The tech debt in our use case involved two steps: Removing the test dependencies and reducing the Docker image size.
I split the cleanup task into two pull requests for Claude Code:
Normally, this would have been a boring, error-prone process, but with AI, it became fast and accurate.
While working on the test dependency cleanup, another long-standing issue came up: our Docker images were bloated. Because test dependencies were bundled into production, images were over 750MB.
With Claude Code, I converted the build into a multi-stage Docker build so only production code was included. The result? A 50% reduction in image size, down to 376MB.
AI agents excel at low-complexity, high-effort engineering work:
This case shows how Claude Code can handle repetitive, time-consuming debt—improving both code quality and developer happiness.
<iframe width="791" height="791" src="https://www.youtube.com/embed/PWpsKdHtsFA" title="How I Tackled Tech Debt with Claude Code - Experience Report" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>
“One of the greatest use cases that I've found for AI agents is to help with technical debt, especially technical debt that is easily fixable, but it just takes a long time to solve. This is the kind of thing that doesn't let you as a developer sleep well at night.
We had in this code base a bunch of test dependencies that were leaking into the production build. And slowly over time, it had grown to about 200 different files with duplicated helper utilities to read JSON files from test resources and that kind of stuff.
And when I started testing AI tools for development, this sounded like the perfect task for it, because it's very easily verifiable, since I'm not touching any logic in the actual code here, I'm just moving test utilities around. As long as my project still builds and passes the tests, we know that we are fine.
So I separated that task into two different PRs. In this first one, I moved only the test utilities for the sources into a separate package and then imported those functions where they were previously used. That resulted in a pretty boring but very accurate PR with 105 files changed. So I did the same thing for the destination utilities in a second PR-–[which was] in total around 200 files that I needed to fix—but of course it was a lot easier with AI.
And the cool thing about this is that it unlocked another thing that was in the back of my head for the longest time, that since we had all of these testing dependencies in the production build, we were including all of that in our Docker images unnecessarily.
So after finishing with the first two, I again use Claude Code to turn my Docker image build process into a multi-stage and as usual, run the dependencies to only include the production code. So that resulted in, as you can see here, in my image when I was testing, around a 50% size reduction in the image. Our images were about 752 megabytes. And after the change, they turned into 376 megabytes.
It was the perfect task for AI because as long as the build and test commands were passing, we know we are good. And for the second task, same thing. And as long as you're done with the Docker build successfully, there is nothing to fear. Everything is fine.”
Ready to reclaim your development time? This case study shows exactly how AI coding agents like Claude Code can transform those lingering tech debt tasks from overwhelming projects into quick wins.
By eliminating 200 tedious file changes in minutes rather than days, we didn't just clean up our codebase—we freed up precious developer hours for the creative, high-impact work that actually moves the needle.
The lesson here isn't that AI will replace developers, but that it can handle the repetitive, time-consuming tasks that keep us from our best work.
Whether it's dependency cleanup, refactoring legacy code, or optimizing build processes, Claude Code turns tech debt from a burden into an opportunity.
The next time you're staring at a backlog of "someday" improvements, consider whether an AI agent might be the perfect tool to finally tackle them—and get back to building what matters.
I publish my thoughts on AI and experience with AI coding tools frequently. Follow me on LinkedIn to stay in touch.
Yandry Perez is a senior software engineer at Faros.
Agent = Model + Harness. Harness engineering is what makes AI agents reliable in production. See the five layers and the metrics that matter.
.avif)