Senior Software Engineer

Preferred candidate profile

Senior Software Engineer - Product Engineer The 15-second version: We build the platform enterprises use to run AI inside their own walls; their cloud, their datacentre, sometimes a network with no internet at all. Banks, insurers, government-adjacent buyers: the people who love what AI can do and are not allowed to send their data anywhere to get it. Were a small team with a product that already runs end-to-end and design partners putting it through real workloads. Early enough that your fingerprints are on the architecture. Late enough that you’re building on something real, not a whiteboard. Why is this job fun? You’ll ship in week one. Not a fortnight of onboarding docs. A real change, in front of real users, with your name on it. No ticket factory. You’ll sit in on customer calls, hear the actual complaint, and go build the fix. The whole loop, not a slice of it. The hard part is genuinely hard. Multi-tenant isolation where a leak is an incident, not a bug. Software that has to install and run correctly in an environment we cannot SSH into. Cost and performance that a customer’s CFO reads every month. Big surface area. Backend most weeks, but you’ll touch frontend, infra and data whenever the problem needs it. Nobody here is the API person. Fast decisions. No design committee, no six-approver RFC. Make the argument well and we do it — often the same day. What you’ll actually do? Own a product area end to end — design it, build it, ship it, and carry the pager for it. Ship to production every week. Watch the dashboards. Fix what you learn. Join customer and design-partner calls. You’ll hear the problem first-hand instead of through three layers of translation. Make real calls: what we build, what we adopt off the shelf, what we delete. Get us into environments most vendors can’t reach — install, run and debug where you can’t just tail the logs. Review other people’s code seriously, and take reviews the same way. What you’ll work with: The real stack, not an aspirational one: Backend: Python, FastAPI, SQLAlchemy Data: PostgreSQL, Redis, pgvector Frontend: TypeScript, React, Next.js Infra: Docker, Kubernetes, Helm, Terraform, AWS. Air-gapped installs are a first-class target, which makes packaging a real problem rather than an afterthought Platform: Temporal for durable work, Keycloak for enterprise SSO, Prometheus and Grafana for knowing what’s happening at 3am AI: LLMs daily, but this is a systems and product job, not a research one. If you want to train models, we’re the wrong shop. If you want to make them cheap, observable and reliable inside a bank, we’re a very good one On Kubernetes: this one isn’t optional. We ship into customers’ clusters, so you need to have genuinely run workloads on K8s in production — not did the tutorial, not our platform team handled it. Come with the war story: what broke, how you found it, what you changed. The rest is flexible. We’ve hired strong people from Go, Java and Rust who were productive inside a month; the language matters far less than the judgment. What are we looking for? Must have: 3+ years shipping software people actually used. Side projects and open source count. Strong in at least one modern backend language, and unafraid of the frontend or the infra when the problem needs it. We hire for judgment, not a keyword list. Real Kubernetes experience — you’ve deployed, debugged and operated workloads on it in production. This is the one hard requirement. You’ve owned something in production. You know what 2am feels like. Solid with a relational database — schema design, and why the query is slow. You can write. A clear PR description and clear thinking are the same skill. You debug from evidence, not vibes. Nice to have: Multi-tenant SaaS, security, or infrastructure-heavy products. Helm and Terraform alongside the Kubernetes — especially shipping software into someone else’s environment. Anything AI/LLM adjacent — genuinely useful, and genuinely not required. Strong engineers pick this up in weeks. You’ve been early somewhere before and know what that actually costs. You’ll love it here if? You’d rather ship something imperfect on Friday than something perfect next quarter. Ambiguity gets you moving instead of stalling you. You have strong opinions and hold them loosely. That’s not my job isn’t in your vocabulary.