AI Engineer / AI Agent Developer

About the role

QNB is growing its AI team and is looking for an engineer with hands-on experience building AI agents. You will work on agentic systems from design to production, together with a team that builds everything in-house.

Key responsibilities

• Design, develop and maintain LLM-based AI agents and multi-agent workflows that automate and enhance internal and customer-facing banking processes.
• Architect end-to-end agentic systems, including orchestration, tool and function calling, memory, retrieval (RAG) and skill/connector integrations.
• Integrate AI agents with internal banking systems, APIs, databases and enterprise data sources in a secure and scalable way.
• Build and maintain guardrails, output validation and safety layers so agent behaviour stays accurate, compliant and aligned with regulatory and data privacy requirements.
• Apply prompt engineering, context engineering and fine-tuning techniques (e.g. LoRA) to optimise model performance for banking use cases.
• Work closely with IT, security and business units to define, prioritise and deliver AI projects.
• Stay up to date with LLMs, agent frameworks and AI tooling, and evaluate new technologies for adoption.

Required profile

• Bachelor's degree from a reputable university, preferably in Computer Engineering, Software Engineering, Mathematics or a related field.
• At least 3-4 years of experience in AI engineering or related roles, including hands-on experience building LLM-based applications.
• Fluency in English, written and spoken.
• Awareness of AI safety, security and data privacy principles in regulated environments.

Required skills

• Strong proficiency in Python and solid software engineering practices.
• LLM APIs and agent frameworks or orchestration tools (e.g. LangChain, LangGraph, LlamaIndex, Langflow).
• RAG architectures, vector databases and embedding models.
• REST APIs and microservices; good command of SQL and enterprise data sources.
• Model serving, inference optimisation or self-hosted open-source models is a plus; speech technologies (STT/TTS) or voice agents is a plus.