Principal AI Software Engineer
Summary
Leads AI-driven software engineering at a fintech prepaid card solutions provider, setting AI coding standards, designing agent-friendly infrastructure, and architecting scalable systems while writing production code and mentoring peers.
About the role
iGoDirect is a leading provider of prepaid card programs and digital payment solutions, enabling businesses to simplify disbursements, incentives, and expense management. We partner with corporates, fintechs, and program managers to deliver scalable, compliant, and innovative payment solutions across Australia.
We are looking for a Principal AI Software Engineer. This is our most senior individual contributor role. You will be technically excellent, implementation focused, and deeply fluent in the way software is built today. Two things define this role. The first is classic principal engineering. You will own architecture, make high impact calls on scale, security and reliability, and write production code alongside the team. The second is new. You will lead how we build software with AI. We treat agentic coding as core engineering practice, not a novelty. Our engineers work with AI coding agents every day. You will not manage people. Your influence will be wide. You will work across multiple projects, guide architecture, unblock engineers, and raise the technical bar through pairing, design reviews and thoughtful code review.
Key responsibilities
Set the standard for how we use AI coding agents across the engineering team
Design and maintain the context layer that agents rely on, including repository instruction files, project conventions, architectural decision records, and machine readable documentation
Build and maintain internal tooling that makes our codebase agent ready, such as MCP servers, custom skills, slash commands, subagent definitions and reusable prompt assets
Define safe operating patterns for agents including sandboxing, permission scopes, secret handling, and human review gates on anything that touches production or customer data
Drive multi agent and parallel workflows where they genuinely help, breaking large migrations and refactors into work that agents can execute in parallel with clear verification
Build the guardrails that make agent output trustworthy through strong typing, high value test coverage, deterministic linting, contract tests, and CI that fails loudly
Evaluate new models, tools and techniques with evidence, running trials, measuring results, and making clear recommendations on what we adopt and what we drop
Coach engineers on effective agent use, including task decomposition, context curation, review discipline, and knowing when to stop delegating and write it yourself
Design and evolve scalable, secure and maintainable systems across our platform, making high impact architectural decisions and communicating the trade offs clearly
Lead by example with clean, high quality code across services and platforms, rapidly prototype, validate and deliver core components and features
About you
10 or more years in software development, with proven full stack depth
Expert level Node.js and TypeScript, with knowledge of the runtime, the async model, memory behaviour and the tooling. NestJS experience is a strong plus
Deep working knowledge of PostgreSQL including schema design, indexing, query planning, transactions, locking and migration strategy at scale
Practical experience with Redis for caching, queues, locks and rate limiting. Familiarity with Valkey, BullMQ or similar job processing is a plus
Hands on production experience with cloud platforms. DigitalOcean is ideal. AWS, GCP or Azure experience transfers well
Daily Docker and containerisation experience. Kubernetes or managed container platform experience is valued
Real experience with OpenTelemetry, structured logging, metrics and tracing, with a track record of using this data to solve genuine production problems
Daily working use of agentic coding tools such as Claude Code, Cursor, Codex, Windsurf, Aider or similar
Ability to describe specific work shipped with agents, including scope, approach, what went well, and where you had to intervene
Experience writing and maintaining agent context, including repository level instruction files, conventions, and documentation written for machine consumption