Lead AI Architect

Role overview

What you will do

• Define AI use cases, user journeys, feasibility, business outcomes, and acceptance criteria with product owners; challenge whether conventional software or analytics is a better fit.
• Lead architecture across experience, application, model, data, integration, and infrastructure layers, documenting decisions, trade-offs, dependencies, and non-functional requirements.
• Design appropriate predictive ML, generative AI, RAG, multimodal, and agentic solutions; use autonomous or multi-agent designs only where valuable.
• Align designs with enterprise architecture, integration patterns, platform standards, and governance; create reusable reference architectures and components.
• Evaluate models, platforms, frameworks, and vendors through structured experiments; recommend build-versus-buy based on quality, security, latency, cost, portability, and operations.
• Guide implementation, architecture and code reviews, integration, and production readiness; prototype high-risk assumptions and mentor engineers and architects.
• Build in privacy, access controls, permission-aware retrieval, tenant isolation, prompt-injection defenses, safe tool use, audit trails, and human approval for high-impact actions.
• Set evaluation, testing, monitoring, and lifecycle controls for models, prompts, retrieval, and agents; plan fallbacks, failure handling, rollbacks, and incident ownership.
• Track AI developments, translate them into practical roadmaps and guidance, and communicate decisions to technical teams, product owners, and senior stakeholders.

Requirements

• Experience leading end-to-end architecture of AI products from discovery through production delivery and continuous improvement.
• Ability to translate business and product goals into secure, scalable, measurable AI solutions.
• Ability to lead hands-on technical architecture, validate design choices with prototypes, and guide teams through delivery.
• Knowledge of predictive machine learning, generative AI, RAG, multimodal models, and agentic workflows to select appropriate solution approaches.
• Ability to address privacy, access controls, permission-aware retrieval, tenant isolation, prompt injection, safe tool use, audit trails, and human approval.
• Ability to define evaluation, testing, monitoring, lifecycle controls, fallbacks, failure handling, rollbacks, and incident ownership.
• Ability to communicate architecture decisions to technical teams, product owners, and senior stakeholders.