AI Platform Engineer (Remote in AU)

Budgetly

1d•Remote

About The Position

This role focuses on building the 'harness' around AI agents, creating an AI-first delivery system where agents ship product features end-to-end. The engineer will build the platform that makes this system safe, reliable, and scalable. The work involves agent runtimes, execution loops, feedback mechanisms, tooling, governance, observability, and evaluation harnesses. Success in this role means delivering a managed agent platform that reliably ships customer-facing features, with engineers focusing on improving the platform rather than direct feature implementation.

Requirements

Strength in systems thinking for agent harnesses and loops.
Ability to design execution harnesses around agents, including feedback loops, evaluation strategy, safety constraints, and 'glue code' for production autonomy.
Strong TypeScript and React experience in production environments.
Experience shipping real software to real users (not just prototypes).
Ability to read a codebase and quickly identify its patterns, conventions, and architecture.
Comfort working in ambiguity and turning fuzzy intent into clear acceptance criteria and evaluations.
Familiarity with agent tooling concepts: tool calling, MCP/tool integration, guardrails, evals, tracing/observability, and permissioning.

Nice To Haves

AWS serverless experience (CDK, Lambda, DynamoDB).

Responsibilities

Design and implement agentic workflows that take a requirement from spec to deploy.
Build agentic loops that turn mistakes into system-level improvements.
Develop evaluation harnesses (offline + CI) to detect regressions in behavior.
Define and maintain review gates (human-in-the-loop + automated reviewers) for risky changes.
Improve tool reliability, including schemas, typed tool interfaces, retries, and timeouts.
Build platform capabilities for managed agents, such as long-running sessions, checkpoints, and recovery.
Evolve the platform architecture with an eye for simplicity and maintainability.
Partner with Product to reduce ambiguity and translate intent into testable, evaluable specs.