Inside Ledgy engineering: Building the future of equity management

Automation is built into every workflow

At Ledgy, we solve complex engineering problems, supporting the equity programmes of some of Europe’s most closely watched, fast-growing companies. In June, we processed more share transactions through our platform than in any other month in our history, reflecting the responsibility our team carries every day.

Because we operate in a highly regulated environment, our engineers have the opportunity to build software where quality and attention to detail directly matter to the companies and shareholders who rely on us.

As the first European equity management platform to launch enterprise-grade AI products, Ledgy is building tools such as Ledgy Agent and Ledgy MCP on top of a modern stack that includes TypeScript, React, Node.js, MongoDB, Temporal, GCP, and Kubernetes. At Ledgy, you’ll help build software that sits at the intersection of automation and modern engineering.

‍

How we migrated a million-line TypeScript codebase

As our codebase grew past one million lines, our developer experience hit a wall. Our TypeScript 6 compiler was constantly running out of memory and taking a long time to check the codebase. We wanted to upgrade to TypeScript 7 to benefit from its native performance, but moving our entire repository to the required “strict” mode felt like an overwhelming, multi-month project.

Rather than treating the migration as a large manual project, we ran a proof of concept using an agentic migration. Using the cost-effective Composer 2.5, we set up an agent to run an overnight loop and work through the migration end-to-end.

Resolving more than 1,000 TypeScript errors required approximately 4,000 lines of code changes at a cost of less than $20 in tokens.

We packaged the proof of concept into a reviewable PR stack. Within one week, the migration was fully deployed. We successfully upgraded to TypeScript 7 with a fivefold speed improvement over TS6, bringing the full type check down to just 30 seconds in CI.

For engineers, this meant removing a major technical bottleneck and creating a stronger foundation with stricter type checks and shorter feedback loops for agentic engineering.

Next in line is an agentic migration from ESLint to Oxlint, to resolve the next major bottleneck in our CI pipeline.

‍

Rewriting over 600 schema contracts without disrupting production

Migrating more than 600 API and database contracts from an old Meteor schema library to the modern TypeScript-first validation library Zod was another major infrastructure milestone. Using a mixture of frontier models such as GPT and Opus and cost-efficient ones like Composer and Grok, we completed the rewrite for approximately $10,000 in model costs without disrupting production.

This rollout was made possible by four testing and review techniques:
‍

• Structural parity testing: We used an agent to write a comprehensive schema introspection toolbox for both libraries. It compared the same schema in both implementations and identified discrepancies
• PRNG-based fuzz testing: Another agent built a deterministic fuzzer generating thousands of random records in our CI pipeline and running them through both schema libraries to identify runtime validation and parsing deviations
• Live validation in production: For several months, we ran both validation layers simultaneously in production. Discrepancies were sent to Sentry, where an automated process translated them into daily pull requests to get schemas more aligned
• Comprehensive reviews: We had long-running agents analyse each schema in the old and new implementation, write comprehensive reports about discrepancies in each, and then resolve them, translating old custom validation code into native Zod features
  ‍

This migration ran quietly alongside day-to-day engineering work throughout the first half of 2026. Once our Grafana dashboard reached 99.9999% agreement between both schema layers, we were able to flip back and forth between the two implementations via a feature flag in production – and nobody noticed.

‍

Join the team building modern equity infrastructure

Ledgy was founded in 2017 by a group of physicists from ETH Zürich who saw a problem first-hand: companies were using equity to attract and retain talent, but the tools available to manage ownership were outdated and difficult to navigate. Since then, we’ve grown into a platform trusted by companies around the world to manage their equity programmes, from incorporation through IPO and beyond.

Today, the engineering team is led by Head of Engineering Angela Santin and Co-Founder and CTO Timo Horstschäfer. Together, our engineers work across product development, infrastructure, security, developer experience, and financial reporting.

This is an exciting time to join Ledgy. You’ll have the opportunity to help shape how modern engineering teams build secure, reliable software for regulated industries, while exploring new approaches to automation and AI.

If that sounds like the kind of challenge you’re looking for, we’d love to hear from you.