How Nubank refactors millions of lines of code to improve engineering efficiency with Devin
Overview
One of Nubank’s most critical, company-wide projects for 2023-2024 was a migration of their core ETL — an 8 year old, multi-million lines of code monolith — to sub-modules. To handle such a large refactor, their only option was a multi-year effort that distributed repetitive refactoring work across over one thousand of their engineers. With Devin, however, this changed: engineers were able to delegate Devin to handle their migrations and achieve a 12x efficiency improvement in terms of engineering hours saved, and over 20x cost savings. Among others, Data, Collections, and Risk business units verified and completed their migrations in weeks instead of months or years.
The Problem
Nubank was born into the tradition of centralized ETL FinServ architectures. To date, the monolith architecture had worked well for Nubank — it enabled the developer autonomy and flexibility that carried them through their hypergrowth phases. After 8 years, however, Nubank’s sheer volume of customer growth, as well as geographic and product expansion beyond their original credit card business, led to an entangled, behemoth ETL with countless cross-dependencies and no clear path to continuing to scale.
For Nubankers, business critical data transformations started taking increasingly long to run, with chains of dependencies as deep as 70 and insufficient formal agreements on who was responsible for maintaining what. As the company continued to grow, it became clear that the ETL would be a primary bottleneck to scale.
Nubank concluded that there was an urgent need to split up their monolithic ETL repository, amassing over 6 million lines of code, into smaller, more flexible sub-modules.
Nubank’s code migration was filled with the monotonous, repetitive work that engineers dread. Moving each data class implementation from one architecture to another while tracing imports correctly, performing multiple delicate refactoring steps, and accounting for any number of edge cases was highly tedious, even to do just once or twice. At Nubank’s scale, however, the total migration scope involved more than 1,000 engineers moving ~100,000 data class implementations over an expected timeline of 18 months.
In a world where engineering resources are scarce, such large-scale migrations and modernizations become massively expensive, time-consuming projects that distract from any engineering team’s core mission: building better products for customers. Unfortunately, this is the reality for many of the world’s largest organizations.
The Decision: an army of Devins to tackle subtasks in parallel
At project outset in 2023, Nubank had no choice but to rely on their engineers to perform code changes manually. Migrating one data class was a highly discretionary task, with multiple variations, edge cases, and ad hoc decision-making — far too complex to be scriptable, but high-volume enough to be a significant manual effort.
Within weeks of Devin’s launch, Nubank identified a clear opportunity to accelerate their refactor at a fraction of the engineering hours. Migration or large refactoring tasks are often fantastic projects for Devin: after investing a small, fixed cost to teach Devin how to approach sub-tasks, Devin can go and complete the migration autonomously. A human is kept in the loop just to manage the project and approve Devin’s changes.
The Solution: Custom ETL Migration Devin
A task of this magnitude, with the vast number of variations that it had, was a ripe opportunity for fine-tuning. The Nubank team helped to collect examples of previous migrations their engineers had done manually, some of which were fed to Devin for fine-tuning. The rest were used to create a benchmark evaluation set. Against this evaluation set, we observed a doubling of Devin’s task completion scores after fine-tuning, as well as a 4x improvement in task speed. Roughly 40 minutes per sub-task dropped to 10, which made the whole migration start to look much cheaper and less time-consuming, allowing the company to devote more energy to new business and new value creation instead.
Devin contributed to its own speed improvements by building itself classical tools and scripts it would later use on the most common, mechanical components of the migration. For instance, detecting the country extension of a data class (either ‘br’, ‘co’, or ‘mx’) based on its file path was a few-step process for each sub-task. Devin’s script automatically turned this into a single step executable — improvements from which added up immensely across all tens of thousands of sub-tasks.
There is also a compounding advantage on Devin’s learning. In the first weeks, it was common to see outstanding errors to fix, or small things Devin wasn’t sure how to solve. But as Devin saw more examples and gained familiarity with the task, it started to avoid rabbit holes more often and find faster solutions to previously-seen errors and edge cases. Much like a human engineer, we observed obvious speed and reliability improvements with every day Devin worked on the migration.
“Devin provided an easy way to reduce the number of engineering hours for the migration, in a way that was more stable and less prone to human error. Rather than engineers having to work across several files and complete an entire migration task 100%, they could just review Devin’s changes, make minor adjustments, then merge their PR”
Jose Carlos Castro, Senior Product Manager
With Devin, Litera cut regression cycles by 90%, delivering new products to customers faster than ever.
About the company
Litera provides trusted solutions to law firms and corporate legal teams worldwide. Their comprehensive suite of AI-driven tools powers and unifies legal workflows across Legal Work & Drafting, Knowledge Management, Business Development, Legal Operations, and Security & Governance.
Overview
With Devin, Litera is collapsing engineering silos, accelerating delivery, and transforming roles—helping the legal tech leader ship more products, faster, and move at the speed of a startup more than ever before.
Litera, a global leader in legal technology, serves thousands of law firms and legal professionals worldwide with solutions that streamline workflows, improve collaboration, and deliver insights at speed. With more than 50 products in market and $300M+ in annual revenue, the company’s rapid growth has brought complexity in engineering with distributed and specialized teams who all needed to respond to growing demands from a competitive marketplace.
The Challenge: Scaling Without Slowing Down
As Litera grew, engineering work fragmented across specialized groups: quality engineers, DevOps, site reliability, and development. Cross-team dependencies delayed bug fixes, testing, and deployments, and QE bottlenecks emerged. With 130 quality engineers stretched across dozens of products, minor fixes could sit in the backlog for weeks. Meanwhile, competitive pressure demanded faster delivery of features and AI-driven enhancements, leaving QE under constant strain.
When CTO Greg Ingino first saw Devin on stage, he immediately recognized its potential: “This is going to be the future of development.” The question wasn’t if they would adopt agent-based development, but how quickly they could make it a reality.
The Approach: QE Evolution and the Path to Autonomous Engineering
Litera partnered with Cognition to launch QE Evolution, an initiative to reimagine quality engineering through autonomous AI agents. Rather than treating Devin as a coding assistant, Litera embedded specialized agents directly into scrum teams, replacing traditional handoffs with autonomous execution.
Working closely with Cognition, the team deconstructed QE functions—test strategy creation, test case development, test execution, automated script creation, and automation triage—then built specialized Devin workflows for each. The goal was to embed specialized skills directly into agile teams and give every engineering manager a “team of Devins” capable of operating as QE testers, SREs, and DevOps specialists simultaneously.
The deployment of new testing agents has resulted in 40% increased test coverage and 93% faster regression cycles. The team now delivers at 10X the output, a change Ingino described as monumental.
Within 30 days, Litera more than doubled active Devin usage, reaching 90+ engineers and growing daily. QE staff began upskilling into agent builders and optimization specialists, ultimately laying the path for transitions into higher-value roles like SRE and DevOps as Devin expanded to other products.
Beyond QE: Scaling Autonomy Across Engineering
While QE Evolution is the most mature initiative, other teams are pushing Devin into new territory. One developer built an auto-scaling Jira system capable of feeding more than ten tasks to Devin at once, dramatically increasing throughput. Agents now validate feature requests and bug reports for completeness before coding begins, ensuring work is properly scoped. Engineers have even used Devin to create a live adoption and productivity dashboard, tracking sessions, pull request volume, and user trends in real time.
Results: Startup Speed at Enterprise Scale
The impact of Devin is being felt across every corner of Litera’s engineering organization. The company launched five major products in just three weeks ahead of its flagship ILTACON event—a pace so fast that competitors have reached out to ask, “What’s going on over there?” Multi-month projects, like 32-bit to 64-bit conversions, now finish in weeks, and high-volume bug triage runs autonomously end-to-end. Some workflows have delivered 5x productivity improvements, enabling what CTO Greg Ingino calls “250 engineers operating at the speed of a thousand.”
The transformation is as much cultural as it is technical. QE staff are moving from manual testing into higher-value roles such as SRE and DevOps, often as agent builders or optimization specialists. One standout example is Volha, who began as a QE individual contributor and, by leading QE Evolution, rose to acting Director of Quality—now overseeing AI agent development and deployment. To further accelerate adoption, Litera introduced a new role “AI composers”—engineers dedicated entirely to solving problems with Devin and other AI tools, parachuting into teams to disrupt timelines and prove what’s possible. “We’ve seen people not just embrace this technology, but use it to completely reinvent their careers,” says Michele Gough, VP of Quality Engineering.
The acceleration is clear to customers. “In my career, I’ve never launched so many products in such a short time,” says Cynthia Gumbert, Chief Marketing Officer. “The innovation pace is unmatched. It’s been fun telling our customers literally every few weeks how much more they can accomplish with Litera’s products.”
This momentum has created what Bill Block, VP, Global AI, Innovation & Architecture, calls the “adoption flywheel”: “Once engineers see they can 10x their output—or stop doing the work they hate—they want in.” Today, Devin is no longer seen as a coding assistant, but as a junior developer, SRE, and QE combined—working autonomously and at scale.
Looking Ahead: Next Steps for Engineering with Devin
For Litera, Devin represents more than productivity gains—it’s enabling organizational reinvention. By embedding autonomous agents into every part of engineering, the company is accelerating delivery, and redefining what’s possible in legal tech. They are now expanding agent integration as part of their PE-backed growth strategy, where rapid innovation and competitive differentiation are critical success factors.
As CTO Greg Ingino puts it: “These results prove that autonomous agents aren’t just tools—they’re team members. We’re not just building software faster; we’re redefining what’s possible when humans and AI work together seamlessly.”
Litera at a Glance
| Company | Litera |
|---|---|
| Industry | Legal Technology |
| Scale | $300M+ revenue, 50+ products, globally distributed engineering teams |
| AI Use Cases | QE Evolution, DevOps automation, SRE agents, requirements validation, Jira auto-scaling, internal analytics dashboards |
| Key Outcomes |
|
| Innovation Approach | Top-down mandate with structured enablement, cross-functional AI champions, and systematic measurement frameworks |