Joe: One month as a software engineer at Octopus Legacy

Before joining Octopus Legacy, I knew nothing about probate. Four weeks later, I’m shipping code that helps people find their way through it. Here’s a look back at my first month.

Off the starting line

I was coming off three years of building in TypeScript and PHP with no commercial Python experience, and my new team was keen to get me up and running quickly. My first task was letting estate planning advisers assign a client case to a colleague without waiting for an admin. That makes a handover, or a reassignment when someone’s off sick, much simpler.

I shipped it on my fourth day. That isn’t unusual here: the engineering team pushes over a dozen changes to production every day, a real step up from the pace I was used to. Matching that pace has been one of my first goals since arriving.

In that same first week, I also removed a retired service, deleting 19,000 lines of code in one go. It’s a power trip I’ll be chasing for a while yet.

I knew nothing about wills or probate before I joined, so there was a lot on the crash-course agenda. After a teammate booked a 1:1 with me, I did the same with the rest of the tech team. It was one of the most valuable things I did to feel settled: I got to know each person better, and learned about different parts of the business from whoever knew them best.

In the deep end with DevOps

After my first contributions, I was drafted in to work on DevOps and AWS, which I’d wanted more exposure to for years. The job was moving a core database onto new infrastructure. While planning the move on staging, I wrote a runbook so the process could be repeated exactly in production, keeping the risk of losing data in the switchover to a minimum.

There’s a lot of trust in the team. When the first change I merged for the move failed, I flagged the omission that had set off the alarm and merged the fix soon after. It’s been affirming to be part of a mature team that trusts its people, even brand-new ones, rather than keeping them in a padded sandbox where nothing they do has consequences.

The magic of BAU Fridays

Plenty of tasks fall through the cracks, and the backlog is where good intentions get forgotten. BAU (business-as-usual) Fridays are how we fight that. We set the last day of the week aside for bug fixes and small feature requests that sit outside the main projects. Without that regular sweep, it would be hard to justify scheduling these important improvements.

One Friday, I built a queue of appointments that needed rebooking, so advisers could track and manage them more easily. It’s a gear shift from my day-to-day work, and I value the change of pace and context at the end of the week.

Reviewing pull requests on a Friday has also helped me get up to speed with the stack. They’re often smaller and need less context, and over time I’ve pieced together knowledge of wide swathes of the business.

Getting stuck into greenfield

I joined almost at the start of our Stealth Fintech Product, a new product built from scratch, so I’ve been heavily involved in scoping it and laying the groundwork. We can’t say much about it yet, but working on something brand new is a challenge I relish: I get a say in how we build it and ownership of my own part of it. Being a go-to person is a satisfying feeling, too.

Scoping as we go has been a challenge (at times it’s felt like the train scene in Wallace & Gromit: The Wrong Trousers, perhaps the best scene in cinema history), but it has made me more independent and taught me to write down my assumptions and share them clearly and early. So far I’ve worked on the technical design, put together wireframes (and been very impressed with the design tool Paper) and built the first screens.

Finding the joy again, with AI

What I’ve always loved about software engineering is the constant micro wins: a working function, a bug fixed, CSS that finally matches a finicky design. AI has largely removed the need for that hands-in-the-code work, which has taken some getting used to. But as I get better at working alongside my agent, I’m finding the enjoyment in the everyday again and growing into more of an architect’s role. The upside is becoming language-agnostic, and being able to focus on the wider problem.

I came to using AI every day later than most, so my biggest challenge has been honing how I use my agent. My first AI-assisted pull requests were too large: hard to digest, and slow to get reviewed. Ones over 500 lines waited about two days for a review, while smaller ones took more like five hours. I’ve since built a harness around my agent that steers it towards smaller, tightly scoped pieces of work that are easier to understand, and my median pull request has dropped from 675 lines to 422.

Not all business as usual

My start date was the week after the summer party, and being invited along was a great chance to meet the team, so I started with familiar faces on day one. The tech team also has a tradition of a lunch in a new joiner’s first week, at a spot they pick. It’s a great way to mark the occasion and help them feel at home early on.

The team is very social in general. We often head out for drinks after work, the company puts on pizza and beer one Thursday a month, and I’ve helped organise a team shuffleboard night. Most importantly, obviously, I’ve been adding to our catalogue of Slack reactions. :martin-lewis: is my personal favourite: we have to give him some love for the regular shoutouts!

Looking ahead after month one

With 30 changes across seven codebases in my first month, I’m pleased with what I’ve achieved in a short time and really motivated to keep progressing with my team.

As the team grows, I’m looking forward to working with more ambitious developers who want to make bereavement easier. I may not be on the phone with families, but I’m proud to play my part in helping Octopus Legacy support people who are thinking about or dealing with death, with the compassion and dignity they deserve.

If that sounds like a team you’d like to join, take a look at our open roles.

Small print: Martin Lewis and MoneySavingExpert aren’t connected with Octopus Legacy and haven’t endorsed this post. The :martin-lewis: emoji is just our team’s own bit of fun.