AI Platform – AI Specialist

We are developing a shared AI platform for the business units of a multi-industry group in Vietnam. The platform runs on Kubernetes with LangGraph, FastAPI, and AWS Bedrock, using LangFuse for observability, and enables business teams to build and operate their own AI agents: attaching tools, loading internal documents, and orchestrating multiple agents that work together.

We’re looking for an AI Specialist to work with business teams, identify valuable AI use cases, and turn them into practical solutions on AI Factory.

You’ll combine business analysis, solution design, and hands-on AI configuration to build assistants and workflows for real users. You’ll help teams choose suitable models, organize knowledge, evaluate output quality, and measure the benefits and operating costs of each solution.

• Lead discovery workshops, analyze current business processes, and assess proposed use cases through the AI Use Case Intake Form, including business value and expected ROI.
• Design future workflows on AI Factory, covering agent architecture, multi-agent orchestration, tools, operating logic, and human approval points.
• Select suitable models for each use case and processing step, balancing cost, speed, and output quality.
• Organize knowledge for RAG, including document sources, collections, and metadata, and define access permissions by role and workspace.
• Build evaluation datasets, metrics, and acceptance thresholds; compare model outputs and measure quality before and after adjustments.
• Define business-specific guardrails and data policies, including restricted content and information that must remain within a workspace.
• Build chatbots and AI assistants, train end users, and incorporate their feedback into ongoing improvements.
• Present implementation results to business leaders, supported by evidence of benefits and operating costs.
• Monitor developments in AI models and tools, and update recommendations for internal use.

• 3–5 years of experience in business analysis, consulting, or solution roles, including at least 1–2 years of direct involvement in AI projects.
• Experience building at least one LLM-based chatbot or AI assistant used by real users.
• A bachelor’s degree in Information Technology, Information Systems, Data Science, or an equivalent qualification.
• Familiarity with major commercial and open-source model families, and the ability to assess their suitability for reasoning, Vietnamese-language tasks, context requirements, speed, and cost.
• An understanding of tool calling, MCP, agent orchestration patterns, and human-in-the-loop workflows.
• Working knowledge of RAG, including chunking, embeddings, hybrid search, reranking, and metadata filtering, with the ability to distinguish retrieval problems from answer-generation problems.
• An understanding of enterprise AI controls, including RBAC, multi-tenancy, workspace isolation, guardrails, audit logs, and personal data protection.
• Practical experience with prompt engineering and LLM evaluation, including reference datasets, metrics, LLM-as-a-judge, model comparisons, and token cost tracking.
• The ability to lead workshops, present to business leaders, and support process changes with evidence. You can explain clearly when a use case does not offer sufficient value to proceed.
• Proficiency with everyday AI tools such as Claude, ChatGPT, Gemini, Microsoft Copilot, and NotebookLM, along with process-mapping tools, Jira, Confluence, and Excel or BI tools.
• Good English reading skills for technical documentation.

Experience designing multi-agent solutions or working on an enterprise AI platform; knowledge of AI governance frameworks; familiarity with n8n, Dify, or Copilot Studio; and industry experience in banking, aviation, consumer finance, or real estate.

We value people who work closely with business teams, stay curious about new AI capabilities, and are honest about results—even when those results fall short of expectations. If you enjoy testing, measuring, and improving solutions until they deliver practical value, we’d like to hear from you.