AI Architect

Booked AI; Melbourne; On-site; AUD 180,000 - 280,000; Full time

Job summary

Booked AI is building AI infrastructure for corporate travel. We seek an AI Architect to take senior ownership of the architecture that makes our agent-driven platform reliable, scalable and easy to use.

You will design, build and improve the systems behind our AI agents, write production code, review architectures and mentor the team. This hands-on leadership role partners with the founder to turn complex travel workflows into a dependable product.

Qualifications

• Proven track record shipping production AI agent systems.
• Experience with orchestration, state management and durable workflows.
• Strong skills in distributed, asynchronous processing and API design.

Responsibilities

• Own architecture and implementation of the AI agent runtime.
• Lead design reviews, code reviews and mentor engineers.
• Ensure secure, reliable, scalable integrations and tooling.

Skills

Agent orchestration

Distributed systems

Python engineering

API design

Architecture leadership

Education

Bachelor's or higher in CS/SE

Tools

Python

FastAPI

PostgreSQL

TypeScript

Job description

The role

Booked AI is building AI infrastructure for corporate travel. Our platform helps travel management companies and their customers search, book and manage travel through natural conversation, with the policies, approvals and controls businesses need.

Our agents work across web, email, messaging and voice, connecting to travel suppliers and business systems to complete real transactions. We are looking for an AI Architect to take senior ownership of the architecture that makes this reliable, scalable and simple to use.

This is a hands-on technical leadership role. You will work closely with our founder and engineering team to design, build and improve the systems behind our AI agents. You will make architectural decisions, write and review production code, and help the team turn complex travel workflows into a dependable product.

The central challenge is making an AI agent act correctly in the real world: understanding a request, using the right tools, maintaining accurate context, respecting permissions and recovering when an external system fails.

What you will own

• The architecture and implementation of our agent runtime, including orchestration, tool use, structured outputs, context management and persistent conversation state.

The boundary between model reasoning and deterministic execution, with clear controls around bookings, payments, approvals and other consequential actions.

Reliable integrations through typed APIs, MCP tools and supplier adapters, including retries, idempotency, reconciliation and recovery from partial failures.

Evaluation and release standards for agent behaviour, using realistic travel scenarios, conversation replays, regression tests and production traces.

Improvements to response time, task completion, concurrency and model cost, guided by measurable outcomes.

Consistent agent behaviour across web, email, Slack, Microsoft Teams, WhatsApp and voice.

• Secure operation across multiple customers, including tenant isolation, role-based permissions, audit trails and protection of traveller data.

• Practical technical guidance for the engineering team through design reviews, code reviews, clear documentation and mentoring.

Our stack

• Backend: Python, FastAPI, Pydantic and asynchronous services.

• AI: LangGraph, Deep Agents, LLM APIs, structured tool calling and MCP/FastMCP.

• Data: PostgreSQL, Redis, SQLAlchemy and persistent agent checkpoints.

Frontend: TypeScript, React, Next.js and generative-ui.

• Infrastructure and quality: Railway, GitHub Actions, LangSmith, Sentry, pytest and Playwright.Integrations: travel distribution systems, airline and hotel APIs, payments, messaging and voice.

You do not need experience with every tool in our stack. Strong engineering judgement and experience delivering reliable AI systems matter most.

What you will bring

• A track record of architecting and shipping production AI agent systems, with direct ownership of architecture, implementation and operational performance.

Deep expertise in agent orchestration, including tool execution, state management, durable workflows, context engineering, memory and recovery across long-running tasks.

• Advanced Python engineering skills and strong foundations in distributed systems, asynchronous processing, concurrency, API design and database architecture.

• Experience designing integration architectures across multiple external platforms, including real-time ingestion, schema normalisation, event processing and reconciliation.

• Proven ability to make agent-driven transactions reliable through typed contracts, idempotency, permission boundaries, human approvals and recovery from partial or ambiguous outcomes.

Experience building evaluation frameworks that measure task completion, tool-use accuracy and failure recovery, with regression testing and production tracing informing release decisions.

• Demonstrated ability to optimise model selection, context size, caching and execution paths against measurable quality, latency and cost targets.

• Strong security judgement across tenant isolation, sensitive data handling, prompt injection and tool access.

• The ability to set technical direction, defend architectural trade-offs, review complex implementations and raise engineering standards while remaining hands-on in the code.

Experience in travel technology, payments, transactional platforms, enterprise SaaS or voice agents would be valuable, but is not essential.

You will have direct influence over a core part of Booked AI’s product and the opportunity to solve demanding problems in AI systems that take real actions for customers. You will work closely with the founder and engineering team, see the impact of your decisions in the product, and help establish the technical foundations for our next stage of growth.