Building real-time insights
for real-world banking
Codat is an advisory intelligence solution purpose-built for modern commercial banking. Through rich, specialized data, forward-looking insights, and integrated workflows, Codat empowers banking teams to deepen their relationships, grow their revenue, and simplify their day-to-day work.
The stack isn't exotic, and that's deliberate
We made choices we can stand behind and invest in improving them over time, like adopting Temporal for durable execution, building a modern data lakehouse, and moving to .NET Aspire so engineers can run complex service dependencies locally without fighting configuration.
The hardest things on the table right now
Our engineering challenges right now span the full stack.
On the platform side, we've just completed a major consolidation: taking a microservice architecture that grew organically and cutting it from ~400 services down to ~100. The next project builds on that: removing our pre-production environment entirely and replacing it with a local developer setup that gives engineers everything they need to work with confidence on their own machines. Beyond a local setup we also have the platform capability to match: automated tests, an observability-first culture, and canary deployments.
On the data side, we're building a modern data lakehouse to power all of our data products, moving from fragmented pipelines to a coherent architecture that can scale. And on the ML side, we're working on Payment Method Inference: extracting payment methods from accounting data where the challenge is that there's almost no standardization in how the same thing gets recorded across different systems. Each iteration gets closer, but the problem space doesn't shrink.
We're also working on making the product available in additional languages, which sounds straightforward until you're actually doing it.
Deliberately light on ceremony,
focused on the work itself
The meetings we keep, we're intentional about. The rest of the time is for focused work, either solo or paired up on a huddle.
Pairing happens naturally. Someone shares a screen, you work through something together, then get back to it. Being remote, huddles do the job that tapping someone on the shoulder used to.
AI tooling is a real part of how people work here. Claude Code is used to parallelize work, not as a novelty. Engineers ship client-facing changes multiple times a day, and there's a weekly town hall to demo and talk through what's been built.
- 01Sprints
- Two weeks. Nothing fancy.
- 02Daily stand-up
- Afternoon, to partner with our US team.
- 03Weekly
- Ceremonies that work for the team.
- 04Roadmap
- Flows bottom-up from the teams doing the work.
- 05Stand-up vibe
- Forgiveness over permission, not status update.
- 06Process
- Teams own their processes and are encouraged to improve them.
Our engineering team in public
How our teams talk about the work in public: customer stories, conference talks, and the occasional deep dive into how the platform actually fits together.
How the Unblocked Context Engine earned a 92 NPS score from Codat engineers
Codat ships AI-native across 200+ microservices. Unblocked is the context engine that hands engineers, and their agents, the full picture, credited in a 30% rise in PR throughput.
Read the story ↗Engineering stories, straight from the team
Deep dives, migrations, honest retrospectives, and what it's like to build here, written by the engineers doing the work.
Read the blog →Our values in engineering
- 01
Own it
Ownership here doesn't stop when your ticket closes. It means chasing a problem all the way to the ground. That might look like spotting rising infrastructure costs and pulling people together to fix it, even if nobody assigned it to you. Other times it means digging into why an incident happened once it's resolved, because "fixed" and "understood" aren't the same thing. We constantly strive to leave things better than we found them.
- 02
Stay curious
We're always looking for a better way, even when the current way is working fine. That curiosity is what led us to adopt Temporal for durable execution, and what gave the platform team a head start on one of our biggest infrastructure projects, because someone decided to test a new tool before they were asked to. It's also what's driven the team to genuinely reinvent how they work around AI tooling, not treat it as optional. If you find yourself asking "is there a completely different way to solve this?", that instinct is valued here.
- 03
Progress over polish
We ship multiple times a day and fix forward when things go wrong. That's not an excuse for sloppiness; it's a deliberate choice about where to spend energy. Each iteration of a model ships when it delivers more value than the last, not when it's theoretically perfect. The discipline is in knowing which trade-offs are worth making and which corners aren't worth cutting. Big-bang projects don't really happen here; progress does.
- 04
You build it, you run it
Engineers own their work end to end. That means thinking about testing, observability, and reliability from the start, not at the point of handoff. There's no QA team to throw things over the wall to, and that's intentional. The people closest to the code are best placed to run it well.
The reasons people stay
- 01Autonomy is real
Engineers from IC1 upwards are expected to identify problems and come up with solutions. If the solution holds up technically, they implement it. No sign-off required.
- 02The culture is no-blame
When things go wrong, and they can, the instinct here is to understand why and fix it, not to find someone to pin it on.
- 03The problems are worth solving
We handle businesses' financial data, which means the work has to be reliable and accurate at scale. There's no shortage of hard problems here.
- 04The ambition is high
A genuine goal to be considered one of the top engineering companies: a place engineers want to work because of how we operate and what we're solving.
What it's like day to day
You'll take on problems that are genuinely hard, work alongside senior engineers who'll get properly into the weeds with you, and be trusted to care about things outside your immediate domain. We have a transparent progression framework, but on a small team, how you show up every day matters more than any box you tick.
IC and engineering management are both genuine paths; no one is pushed into management because it's the only way to grow.
Technical strength matters here, but so does everything else
The engineers who do best here are product-minded, proactive about finding and fixing problems, and comfortable with end-to-end ownership. They have strong opinions and argue for them, and commit once a decision is made. They embrace ambiguity rather than waiting for it to be resolved for them.
- Product-mindedEven when the customer is internal.
- ProactiveFinds and fixes problems before being asked.
- End-to-endFrom writing the code to running it in production.
- OpinionatedStrong views: willing to argue, willing to commit.
- Comfortable with ambiguityDoesn't wait for it to be resolved.
- AI-fluentNot a gimmick, but part of how they work.
Designed to move at speed
Talent screen
An intro to Codat, and to establish whether it's worth going forward. No prep needed. We'll ask about your background and what's drawing you to a role at Codat.