Yesterday I was told that, after three years, my position is being eliminated and that I'll keep working through the end of September. These moments are never easy, even less so living in another country and being the main income for my family.
But I'd rather write this from where I actually stand: calm, and optimistic.
What's next
I'm starting the process of finding a new software developer role again. I have until September, a good runway, and I'm confident I'll find something good within that window.
If you know of an opportunity or just want to talk, reach out. I'm open to both full-time roles and freelance work.
Grateful for these three years
I'm deeply grateful for my time at what is still my company. They always treated me well and I have nothing but good things to say, only gratitude.
Being a startup, I learned a huge amount about technologies I'd never worked with before: Next.js, React, MongoDB, ClickHouse, Blazor and Azure Functions, among others. A few of the things I built there:
- Share of Wallet — a brand-new survey question type I built from scratch, end to end. It lets a golf club ask its members which other clubs they also play at and how often, then turns those answers into a chart showing what share of their golf happens elsewhere. I built the whole stack: the data model, a geo-radius club search (finding nearby clubs within a set distance in km² using MongoDB's geospatial queries), the API endpoints, the internal Blazor editor staff use to configure the question, and the ClickHouse pipeline that aggregates raw answers into the final dashboard chart.
- Natural-language queries over each club's data, powered by an LLM.
- Retention Intelligence — a new product page on the frontend that flags members at risk of leaving, with KPI cards, charts, a churn-risk table, and revenue-at-risk figures, plus an AI that takes all of that information and returns a summary of what's shown in no more than 150 characters (Next.js, React, TanStack Query/Table, charts).
- A lot of work on the company's CRM. Each change to a "parent" club could be propagated (or not) to its "child" clubs. To do that, I first had to check whether certain data was exactly the same between each child and the parent: based on that, we could propagate changes only to the children that matched the parent, or to all of them. I also had to surface which clubs had differences and what those differences were. The comparison ran by section, depending on what was being edited, so we could be looking for differences in anything from a questionnaire with multiple options, where a question only shows up if a previous one was answered a certain way, to newly added languages that a child might not have yet, among other things. It was a challenging and genuinely interesting thing to build.
The upside
I'll have more time to finish my main project, ImIn. So expect more frequent updates here.
If these years taught me anything, it's that building is the best way to keep moving. See you in the next post.