Agentic AI Architect / Generative AI Architect

A major US multinational is expanding its Data & AI consulting practice out of Singapore. They are hiring a practitioner-architect to be the technical owner for all client AI deliveries.
You will be the most senior hands-on architect in the team and the go-to technical face for clients. Your job is to decide how we build - when to go agent-first and when to keep it simple and deterministic - and to make sure what we build is secure, governed, production-ready and defensible to risk, security and audit teams.

You will work with a small build team in Singapore [AI Engineers, Data Engineer, Data Scientists], partner with global Data & AI and Controls teams, and sit across the table from CTOs, CDOs, Enterprise Architects and CISOs/CROs.

What You'll Do

1. Lead Architecture for Client Builds

Define the target architecture from problem framing to live production. Convert business use-cases into clear blueprints - components, data flow, model choices, integration approach, security and governance. Create reusable reference architectures for both enterprise AI and agentic systems - multi-agent workflows, A2A and MCP tool layers, gateway patterns, RAG / knowledge systems, and enterprise integration. Set the standard for when agentic actually helps vs when it doesn't.

1. Own Platform, Model & Risk Decisions

Own the model/vendor strategy. Evaluate and compare frontier closed models and open-weight models on performance, cost, latency, context, residency, security and portability. Build evaluation criteria, selection guardrails and exit strategies so clients don't get locked in. You own the build-vs-buy call.

1. Bake in Security, Controls & Governance

Design with controls on day zero - injection safeguards, least-privilege agent identities, tool permissioning, sandboxing, human-in-the-loop, DLP, memory integrity, sovereignty, audit trails and evidence. Make sure designs can pass muster with regulated clients, boards and regulators across APAC.

1. Set the Technical Bar for the Practice

Define architecture principles, patterns, NFRs, and design review / assurance rituals. Review engineering work, mentor the team, and raise overall quality. Partner with practice leadership to shape proposals - technical approach, delivery plan, risk stance - and build reusable assets, accelerators and methods.

What We're Looking For
6-10 years in solution / enterprise / data / cloud / AI architecture or consulting, with real experience shipping production AI / data / platform systems.
Proven track record building production GenAI and agentic systems end-to-end - orchestration, retrieval, model integration, enterprise systems, and moving from POC to prod.
Comfortable as the client-facing technical lead - running architecture workshops with senior enterprise stakeholders.

Strong Advantage If You Have:
Deep agentic expertise - multi-agent orchestration, A2A, MCP, gateways, RAG, vector stores, knowledge graphs, evals.
Hands-on with frameworks like LangGraph, CrewAI, AutoGen, Microsoft Agent Framework or similar.
Good understanding of model landscape - Anthropic / OpenAI / Gemini + Llama / Gemma / Phi / Mistral / Cohere and how to pick between them.
Exposure to APAC models - Qwen, DeepSeek, GLM, Kimi, Yi, HyperCLOVA X, EXAONE, ELYZA, SEA-LION etc.
Strong on cloud AI stacks - Azure [Foundry / AOAI], AWS Bedrock, GCP Vertex - plus knowledge of sovereign clouds for residency - Alibaba, Tencent, Huawei, Naver, self-hosted inference.
Solid integration design - APIs, events, identity, enterprise apps, data platforms - plus MLOps / Lakehouse / observability.
Security for AI systems - injection defense, tool auth, identity, approvals, audit, privacy