Assistant Manager, Information Technology (AI Application Development)

Responsibilities:

• Collaborate with business and IT teams to understand workflows and requirements.
• Prototype, validate, and build AI, machine-learning, and automation solutions.
• Assist in scoping integrations with existing applications, databases, APIs, identity services, and infrastructure, identifying risks and dependencies early.
• Define measurable acceptance criteria, evaluation datasets, rollout approaches, and support arrangements in collaboration with project stakeholders.
• Design, develop, and implement end-to-end AI/ML solutions, including predictive models, NLP, computer vision, decision support, automation, and generative AI applications.
• Build and integrate robust generative AI patterns (such as retrieval-augmented generation (RAG), fine-tuning, prompt engineering, and guardrails) and implement LLMOps practices for model evaluation, monitoring, and safe deployment.
• Prepare and process data for AI solutions, including feature engineering, labeling, embeddings, vectorization, and data-quality checks, to build reliable and maintainable production components.
• Support the deployment of solutions through controlled pilots and production rollouts and troubleshoot incidents.
• Monitor quality, latency, cost, security, adoption, and business outcomes.
• Implement AI strategy, policies, standards, and technical controls (including security, privacy, access-control, and Responsible AI practices) within assigned solutions.
• Assist in evaluating and experimenting with emerging AI technologies, including foundation and multimodal models, agentic workflows, vector databases, and orchestration frameworks to build focused proofs of concept.

Requirements:

• Bachelor’s degree in Computer Science, Artificial Intelligence, Data Science, Statistics, Engineering, or a related quantitative field.
• Hands-on experience in AI/ML, data science, or generative AI development (including prototyping, building, integration, or deployment), with around 5 years of relevant IT experience.
• Proficiency with Python and common AI/ML frameworks and tools (e.g., scikit-learn, PyTorch/TensorFlow, Hugging Face Transformers, LangChain/LlamaIndex).
• Experience in designing and implementing LLM solutions, including RAG pipelines, prompt engineering, fine-tuning, and LLMOps, with an understanding of cloud-based, on-premises, or hybrid deployments.
• Good understanding of security, privacy, and compliance requirements for AI (PII handling, access controls, encryption, secrets management, prompt-injection defenses).
• Strong analytical, systems-thinking, troubleshooting, and delivery skills, with a track record of delivering production-ready AI or data solutions.
• Excellent communication and interpersonal skills, with the ability to explain complex technical concepts clearly in both English and Chinese to diverse stakeholders at all levels.
• Proactive, innovative, security-conscious, and comfortable working independently in a dynamic, rapidly changing technical environment.
• Excellent command of written and spoken English and Chinese.

Closing date for application:

24 September 2026

How to Apply?

Please click the "Apply" button. Full resume with present and expected salaries should be submitted via online application at https://careers.ura.org.hk/.

For more information on the work of the URA, please visit the website at http://www.ura.org.hk.

All applications will be treated in the strictest confidence and information provided will be used for the consideration of applications. Applicants not invited for interview within four weeks from the closing date may consider their applications unsuccessful.

Personal data provided by job applicants will be treated in strict confidence and used strictly in accordance with our “Personal Information Collection Statement Pertaining to Recruitment” for recruitment purpose only. A copy of the Statement is available on our website and obtainable on request. Applicants may be considered for other suitable positions in the Authority and all personal data of unsuccessful applicants will be retained for a maximum of two years.