Principal AI Architect

About the role

We are building a governed AI function across a large private hospital group in Indonesia. This role is the technical quality gate for everything we deploy. You will hold formal production clearance authority over all AI initiatives in the division — vendor-sourced, open-source, and internally built. Nothing ships without your sign-off. You will execute and adapt an evaluation framework set by the line manager, mentor a junior engineering team toward higher standards, and advise on build-vs-buy decisions with documented reasoning. This is a judgment role, not a delivery role. The person we are looking for evaluates, governs, and enforces — and has the credibility and presence to make engineers want to meet the standard, not merely comply with it.

What you will do:

Evaluate

• Assess AI solutions — vendor, open-source, cloud, and internal — against criteria covering performance, reliability, interpretability, clinical appropriateness, data pipeline integrity, and integration feasibility.
• Lead build-vs-buy technical evaluation for all AI components, with written recommendations and documented reasoning.
• Conduct technical due diligence on vendor claims, including where benchmarks are insufficient or non-transferable to our clinical population.
• Issue written evaluation reports that are technically precise and accessible to non-technical stakeholders. Govern
• Execute and adapt the division's AI Evaluation Framework across the full portfolio.
• Maintain a live register of AI components: evaluation status, production readiness, monitoring state.
• Set engineering standards for AI work produced within the division: code quality, model documentation, evaluation methodology, handoff protocols.
• Identify technical debt in the current AI project portfolio and propose a prioritised remediation plan.

Mentor

• Mentor engineers toward competencies in AI evaluation, integration engineering, and responsible AI practice.
• Assess team capability honestly and constructively, identifying individuals with genuine high potential.
• Participate in directorate-facing engagements — Community of Practice, pilot discussions — as a credible, non-partisan technical voice.

What we need from you

Experience

• Highly experienced across AI/ML, systems architecture, and regulated-industry or healthcare contexts.
• Comfortable wearing at least three of four hats: AI/ML technical judgment, systems architecture, project governance, healthcare domain knowledge. No single hat is sufficient.
• Demonstrable experience evaluating AI systems in production — not just building them. You can name failure modes, not just metrics.
• At least one example of making a governance framework operational in an organisation where compliance had previously been inconsistent.
• Healthcare or clinical operations exposure strongly preferred.

Technical

• ML model evaluation: metrics, statistical validity, bias assessment, interpretability methods, and the limits of benchmark performance.
• LLM capabilities and failure modes: where they are appropriate, where they are oversold, and how to stress-test a vendor's claim in a clinical context.
• Sufficient systems architecture literacy to assess integration risk — you do not need to build pipelines, but you need to know when one will fail.
• Familiarity with AI governance frameworks and emerging regulatory expectations for healthcare AI.