AI Engineer

Role Summary

UltreonAI is an AI Design Studio, at the forefront of designing and deploying custom Agentic AI solutions specifically tailored to enhance the efficiency and competitiveness of Singaporean SMEs. Our mission is to move businesses beyond traditional systems by implementing smart, context-aware AI agents.

This role offers a unique opportunity to be a foundational part of this mission. The AI Engineer will be responsible for architecting and implementing the intelligence layer of our custom AI solutions. You will design AI systems that go beyond simple API calls—building sophisticated RAG architectures, orchestrating multi-step agent workflows, and ensuring AI outputs are reliable, cost-effective, and aligned with client business requirements.

Your primary focus will be on designing, implementing, and optimizing the AI systems that power our client solutions.

• AI System Architecture: Design and implement complete AI systems—not just API integrations, but thoughtful architectures that handle retrieval, reasoning, tool execution, and response generation in production environments.
• RAG Pipeline Development: Build and optimize Retrieval-Augmented Generation systems including document processing, chunking strategies, embedding generation, vector search optimization, and context assembly for LLM prompts.
• Prompt Engineering: Develop robust, structured prompts that reliably produce desired outputs across diverse use cases. Implement prompt templates, few-shot examples, and structured output schemas to ensure consistent AI behavior.
• Agent Orchestration: Design multi-step agentic workflows where AI agents make decisions, execute tools (database queries, API calls, web searches), manage state, and handle complex business logic autonomously.
• Quality Assurance & Evaluation: Establish evaluation frameworks to measure AI output quality, identify failure modes, detect hallucinations, and implement mitigation strategies. Monitor system performance and iterate on improvements.
• Cost & Performance Optimization: Manage token usage, implement caching strategies, optimize context windows, and balance quality vs cost tradeoffs across different LLM models and deployment strategies.

• AI System Design: Deep understanding of how to architect AI systems—not just calling LLM APIs, but designing complete pipelines including retrieval, context preparation, prompt orchestration, output parsing, and error handling.
• RAG Architecture: Hands-on experience building Retrieval-Augmented Generation systems. Understanding of chunking strategies, embedding models, vector similarity search, hybrid search approaches, and context ranking/reranking techniques.
• Prompt Engineering: Expert-level prompt engineering skills—structured outputs (JSON mode, function calling), chain-of-thought prompting, few-shot learning, system message design, and prompt optimization for reliability and cost.
• Agent Orchestration: Experience designing agentic workflows where AI systems make multi-step decisions, use tools, maintain conversation state, and handle complex reasoning tasks. Familiarity with frameworks like Lang Chain, Llama Index, or custom orchestration logic.
• Type Script & Node.js: Strong proficiency in Type Script for AI integration work. Comfortable building AI pipelines in Node.js environments and integrating with backend APIs.
• LLM APIs: Deep familiarity with LLM APIs (OpenAI, Anthropic, or equivalent)—streaming responses, function calling, token management, error handling, rate limiting, and API cost optimization.
• Vector Databases: Hands-on experience with vector databases (Pinecone, Weaviate, Chroma
  
  DB) for semantic search, including index optimization, metadata filtering, and hybrid search implementations.
• Redis: Understanding of using Redis for AI context caching, conversation memory, rate limiting, and managing stateful agent interactions.
• Evaluation & Quality Control: Ability to design evaluation frameworks—measuring accuracy, relevance, coherence, and detecting failure modes like hallucinations or off-topic responses.

• Data Preparation: Workin…