Director of AI

Direct Travel•Remote,

8h•Remote

About The Position

Direct Travel is building artificial intelligence into how we operate and what we deliver to our clients. Our objective is what we call the perfect trip — a travel experience where intelligent systems anticipate needs and remove friction at every step. The Director of AI is instrumental to delivering this strategic vision and roadmap. The central mandate is agentic. Leveraging our data foundation, you will enable intelligent agents to drive business processes end to end — establishing the AI platforms and frameworks the organization builds on, driving the engineering of the harnesses those agents operate within, and measuring rigorously whether they work. This is a hands-on leadership role. You will own an AI budget, a global team of full-time and contract engineers, and a vendor portfolio, while remaining active in architecture and design decisions. Technical credibility with the team you lead is essential.

Requirements

Bachelor's degree in Computer Science, Artificial Intelligence, Data Science, Engineering, Mathematics, or a related technical discipline required.
10+ years in software engineering leadership, including 5+ leading AI transformation at organizational scale — not isolated pilots.
Experience leading global teams that mix full-time employees and contract engineers.
A demonstrable track record of using AI to scale organizational output.
Experience establishing an AI-driven engineering culture, and owning a budget including vendor spend and build-versus-buy decisions.
Strong engineering fundamentals and current hands-on credibility — you contribute to architecture, design, and implementation rather than only reviewing them.
Solid command of the software development lifecycle, source control, CI/CD, and production operations.
Have personally built and shipped production AI systems, including agentic ones, and can account for how they were architected, integrated, and evaluated.
Practical experience implementing retrieval-augmented generation (RAG) in production, including its failure modes.
Fluency in agent integration — MCP, A2A, or equivalents you have built yourself — and in evaluation practice: measuring quality, catching regressions, and knowing when something is ready.
Current understanding of the model landscape — frontier and open-weight models, their tradeoffs, and how fast it changes.
Fluency across major cloud providers and their AI services, and command of AI unit economics: inference cost, model routing, and self-hosting tradeoffs.
Working understanding of the AI regulatory landscape, including ISO/IEC 42001 and the EU AI Act, and what they require in practice.
Experience delivering AI inside enterprise governance, cybersecurity, and data privacy constraints, and understanding of AI-specific security risks.
Able to explain AI capability and limitation honestly to executives, clients, and engineers.
Comfortable making principled decisions in a field that outpaces written requirements.
Willing to say when AI is the wrong tool.

Nice To Haves

Master's degree in Artificial Intelligence, Machine Learning, Computer Science, Data Science, or a related field preferred. Equivalent combination of education and directly relevant experience will be considered.
Hands-on experience building ML models from the ground up — training, tuning, and deployment, not API consumption alone.
A data science background sufficient to interrogate model behavior and evaluation results.
Experience in travel, hospitality, or logistics, particularly at global or multi-brand scale.
Experience with workflow automation platforms, particularly n8n.
Experience operating AI capabilities in multi-tenant environments serving external clients.
Experience taking an organization through ISO/IEC 42001 certification.

Responsibilities

Define and deliver the agentic AI strategy — intelligent agents operating on our consolidated data to drive business processes end to end.
Drive the engineering of the harnesses these agents run inside — tool access, context management, control flow, error recovery, and guardrails — and set integration standards including the Model Context Protocol (MCP) and Agent2Agent (A2A).
Build the evaluation discipline that proves an agent works: automated evaluations, regression testing, and clear criteria for promoting from pilot to production.
Define where humans stay in the loop, particularly where decisions carry financial and service consequences.
Establish the AI platforms, frameworks, and reference architectures the organization builds on, so capability compounds rather than being rebuilt project by project.
Own model strategy: which providers, which models for which workloads, and when a smaller or self-hosted model beats a frontier one.
Own AI unit economics — inference cost, cost per workflow, and the decisions that keep spend proportionate to value.
Build AI into the services and products our clients use, in direct pursuit of the perfect trip.
Own the AI capabilities embedded in travel operations — vendor platforms where they fit, built in-house where they do not.
Raise organizational output through AI — internal tools, chatbots, copilots, and automation people actually adopt — and prioritize what is worth funding.
Lead AI literacy across the organization, building genuine capability in the teams using these tools daily.
Drive adoption against the ordinary friction of change.
Continue establishing an AI-driven engineering culture, where building with AI is normal practice and intentional.
Ensure AI complies with enterprise governance, cybersecurity, data privacy, and regulatory standards, partnering with security, legal, and privacy so it is built into how we ship.
Build toward ISO/IEC 42001 alignment and EU AI Act readiness, including risk classification of AI use cases and the supporting documentation.
Own AI-specific security posture: prompt injection, exfiltration through agent tool access, over-permissioned agents, and the review process that catches them.
Lead and grow a global team of full-time and contract engineers, and own the AI budget across headcount, vendor platforms, infrastructure, and model spend.
Own build-versus-buy judgment. We run a select portfolio of vendor AI platforms; you drive what stays, what gets replaced, and what we build ourselves with engineering leadership.
Manage vendor relationships as a core function: evaluation, negotiation, performance management, and knowing when to walk away.