Software Engineer

About Purelend

Purelend is an early-stage company operating at the frontier of FinTech and AI. Our small, tight-knit founding team is reimagining banking technology, starting with mortgage loans. We're fully automating the manual checks and underwriting which historically have meant hours or days of frustrating manual work. Our customers and their clients get what they want faster: a funded mortgage.

Our platform is already used by mortgage brokers and lenders across Canada, including some of the country's most important lenders. Revenue is growing, the pipeline is full, and every line of code you write moves a real number on the business.

Our AI Stance

We're all in on AI, and not just in the product. Engineering, infrastructure, design, customer support, sales: every function here uses the best AI tooling we can find to maximize quality and throughput on a small team. We're looking for someone who wants to help us push that bet further, not someone who's skeptical of it.

The Role

We're hiring a Software Engineer to join our core product team and own meaningful chunks of the platform end-to-end. This isn't a role where you get handed a Jira ticket and a fully-specced design. You'll talk to customers, scope your own work, write the code, ship it, watch the metrics, and own what you built.

The stack is Next.js, TypeScript, Node.js, and Postgres, with AI and LLMs woven through the product wherever they earn their keep. You'll work across the stack: backend services, database schemas, AI evaluation pipelines, and the customer-facing UI brokers and lenders use every day.

What You'll Actually Do

Ship Features End-to-End

You'll take a problem like “brokers can't tell at a glance which large deposits in a bank statement need sourcing” and turn it into a shipped feature. That includes the data model, the API, the LLM prompt or extraction pipeline, the UI, the tests, and the observability. No handoffs.

Productionize AI

A lot of what we build relies on LLMs doing useful work reliably: classifying documents, extracting income from bank statements, flagging suspicious patterns, summarizing files for underwriters. You'll design the prompts, build the evals, wire up the fallbacks, and decide when a deterministic rule beats a model call. We expect rigour here. Demos are easy. Reliable AI in production is the bar.

Own the Data Layer

Postgres is the source of truth for almost everything we do. You'll design schemas you'll be happy to look at in two years, write queries that don't fall over under real load, and care about migrations enough not to break things on the way through.

Talk to Customers

You'll be on calls with brokers and lenders, often alongside the founders. Sometimes you're demoing what shipped this week. Sometimes you're digging into a bug with them on a screenshare. Sometimes you're asking three questions that turn into next quarter's roadmap. If the idea of getting on a call with a customer makes you tense up, this isn't the right place.

Raise the Bar on the Codebase

You'll review code, push back on shortcuts that aren't worth taking, and tighten the parts of the system that hurt most. We move fast, but we're building infrastructure that financial institutions depend on. We don't ship junk and hope.

What You Bring

• 3–7 years of professional experience building production software. You've owned features in production long enough to feel the difference between code that works on your laptop and code that survives real users.
• Strong TypeScript across the stack. You can read and write Next.js, server actions, API routes, and Node.js services without reaching for the docs every five minutes. You understand the runtime, not just the syntax.
• Postgres fluency. You can design a normalised schema, write a join you're not ashamed of, and read an EXPLAIN plan well enough to know when an index is the answer and when it isn't.
• Hands-on AI / LLM experience. You've shipped at least one real LLM feature in production. You know what evaluation looks like (Evalite, Braintrust, custom harnesses, whatever), and you've been burned at least once by a model that drifted on you. Bonus if you've worked with the Vercel AI SDK, LangChain, structured output, or function calling in anger.
• Comfort with the full lifecycle. You can deploy, monitor, and debug a service in production. You don't need a separate DevOps team to ship a feature.
• Customer-facing comfort. You can sit on a call with a head of underwriting at a national lender, ask the right questions, and translate what you hear into a roadmap item.
• Taste. You care what the product looks like and how it feels to use. You'll push back when a design or interaction isn't good enough, and you'll have a real opinion about how to fix it.
• An AI-first mindset for your own work. You use the best AI tooling available to compound your output: coding agents, evals, internal automation, scripts that didn't exist last week. You treat AI as part of the craft, not a threat to it.

What We're Not Looking For

• Engineers who need every requirement specced out before they can start.
• People who think prompts are the whole job. LLM features get hard exactly when you go past the demo.
• Anyone who's allergic to talking to customers or shipping something that isn't perfect yet.

Why This Role Is Interesting

You're joining at the inflection point. Product works, revenue is growing, and the institutional deals in the pipeline are the kind that change the trajectory of a company. The engineering team is small. What you ship in your first quarter will still be in production years from now, touched by every customer.

Compensation

Our total compensation, a blend of cash and equity, is market competitive.

Details

• Location: Calgary, AB — in-office
• Reports to: Seb Hiscock (CTO)