Applied AI Engineer – LLM & Agentic Systems

• CompanyDION – Digital Innovation
• Working modelRemote
• EmploymentFull-time
• CompensationUp to 35,000,000 VND/month
• LocationVietnam
• EnglishRequired

About the role

DION is looking for an Applied AI Engineer to join our AI engineering team and build production-grade AI applications for enterprise customers.

This role focuses on the engineering and implementation of LLM-powered systems, AI Agents, RAG pipelines, workflow automation and backend integrations.

You will work on real-world AI systems rather than research prototypes, with particular attention to reliability, structured outputs, state management, evaluation, observability and integration with enterprise systems.

Responsibilities

• Design and develop LLM-powered applications and autonomous AI Agent workflows.
• Build stateful and multi-step agent workflows using LangGraph, LangChain, CrewAI or equivalent frameworks.
• Implement tool and function calling, agent routing, multi-agent orchestration, context management, checkpointing, workflow resumption and human-in-the-loop workflows.
• Build RAG and enterprise knowledge systems using vector databases such as pgvector, Qdrant and Pinecone.
• Design document ingestion, chunking, embedding, retrieval, reranking and context optimization pipelines.
• Develop robust structured-output pipelines using Pydantic, JSON Schema and deterministic validation.
• Prevent common LLM production issues such as malformed output, hallucination, inconsistent responses and workflow failures.
• Build backend services using Python, FastAPI and asyncio.
• Integrate AI systems with enterprise APIs, databases, webhooks and third-party services.
• Implement streaming, background processing, retries, timeout handling, rate limiting and secure tool execution.
• Build automated LLM evaluation and regression testing pipelines.
• Evaluate AI systems based on accuracy, hallucination rate, retrieval quality, latency, token usage and API cost.
• Implement monitoring and observability using LangSmith, Langfuse, DeepEval, MLflow or equivalent solutions.
• Work closely with technical leads and international project teams to translate requirements into reliable production systems.
• Review technical solutions, troubleshoot production issues and continuously improve system quality.

Requirements

Must have

• 3+ years of professional software development experience, with strong experience in Python backend development.
• Strong knowledge of Python, FastAPI or equivalent backend frameworks, asyncio, REST APIs and type annotations.
• Practical experience building LLM applications beyond basic prompt or API wrappers.
• Hands-on experience with at least one AI agent framework such as LangGraph, LangChain, CrewAI or a custom agent/state-machine implementation.
• Good understanding of agent state management, tool calling, function execution, agent routing, multi-step workflows and structured outputs.
• Experience developing RAG systems.
• Experience with at least one vector database such as pgvector, Qdrant, Pinecone or equivalent.
• Experience using Pydantic and JSON Schema for structured data validation.
• Good knowledge of PostgreSQL and Redis.
• Experience with Docker and Git.
• Ability to write maintainable, testable and production-quality code.
• Good spoken and written English, including the ability to read technical specifications, communicate technical issues and participate in meetings with international teams.

Highly preferred

Experience with one or more of the following is a strong advantage:

• LangSmith, Langfuse, DeepEval, LLM-as-a-Judge evaluation or synthetic test dataset generation.
• Instructor, Outlines, OpenAI Structured Outputs or other strict-schema approaches.
• Hybrid search, reranking, context compression or context pruning.
• Prompt and model evaluation across OpenAI, Anthropic Claude, Gemini or open-source LLMs.
• AWS, Azure or Google Cloud; Kubernetes and CI/CD.
• MLflow, OpenTelemetry, AI security and guardrails.
• Multi-tenant enterprise systems.

What we are looking for

We are particularly interested in engineers who understand that building production AI systems is not only about prompting an LLM. A strong candidate should be comfortable answering questions such as:

• How do you resume an interrupted AI Agent workflow?
• How do you prevent an LLM from returning invalid JSON?
• How do you validate an AI output against business rules?
• How do you evaluate whether a new prompt or model version is actually better?
• How do you detect and measure hallucinations?
• How do you control token cost and latency?
• How do you design reliable tool-calling workflows?
• How do you build RAG systems that return evidence instead of simply generating plausible answers?

What we offer

• Salary up to 35,000,000 VND/month, depending on experience and technical capability.
• Remote working.
• Opportunity to work on real enterprise AI projects.
• Hands-on exposure to modern LLM and Agentic AI technologies.
• Work with production systems involving AI Agents, RAG, Document AI, enterprise automation and workflow intelligence.
• Opportunity to work directly with international customers and technical teams.
• Strong opportunities to grow toward Senior Applied AI Engineer, AI Tech Lead or AI Solution Architect.