Build AI that ships, not AI that demos.
Lion Rock Technology has been building software in Hong Kong since 2018. Today that means agentic AI, cloud platforms, apps and web systems for local clients — plus our own products, CleverKeeper and GymRock, which run in production with real users. If you would rather own an outcome than sit in a delivery layer, read on.
Four things that shape the job.
We describe the work, not the perks. These are the conditions you would actually be working in.
AI-native delivery
AI is in the toolchain, not in the pitch deck. Everyone here uses coding agents, LLM pipelines and AI-assisted design as normal working tools, and we expect you to argue about where they help and where they do not.
Small pods, real ownership
Projects run with a handful of people and one named owner. You will talk to the client directly, make the call, and carry it through to acceptance — there is no account-manager relay to hide behind.
Shipped products, not slideware
CleverKeeper and GymRock are live products we designed, built and still operate. What you learn from running something in production goes straight back into client work, and the reverse.
Hong Kong clients, real constraints
Bilingual interfaces, tight budgets, NGO funding cycles, systems that must keep running for years. The constraints are the interesting part, and they are why the work stays concrete.
Eight seats, all in Hong Kong.
Every role is based in Hong Kong and works with the same clients and the same codebases. Apply by email — tell us which role and what you have built.
Agentic AI Engineer
You build LLM agents and tool-use pipelines that take real actions in production — booking, billing, reading documents, updating systems of record. That means designing the tools, the guardrails and the evaluation harness, then keeping the thing honest once real users are on it.
Apply for this roleSenior Cloud Engineer
You architect and run the cloud infrastructure behind client systems: networks, deployments, backups, monitoring and the security posture that holds them together. You also own the bill, so you will be the one asking whether a workload really needs what it is asking for.
Apply for this roleFull-Stack Engineer, Web & Mobile
You ship client web platforms and mobile apps end to end — schema, API, interface, store release. Most of our work is bilingual and has to stay maintainable for years, so we care more about clear code you can hand over than clever code you cannot.
Apply for this roleAI Adoption Specialist
You help client teams actually put AI into their daily operations — sitting with the people who do the work, finding the tasks worth automating, and building the workflow with them. Success is measured by what the team still uses three months later, not by the workshop.
Apply for this roleSenior UI/UX Designer
You take a product from user research through to a production design system for apps and platforms — flows, states, edge cases, the Chinese and English versions of every screen. You work alongside the engineers who build it, so your files have to survive contact with real code.
Apply for this roleProduct Designer, AI Prototyping
You set the art direction for a concept and then prove it fast, using AI tooling to get from a rough idea to something clickable in days. Half the job is taste; the other half is knowing which prototypes to throw away before anyone gets attached to them.
Apply for this roleAssistant Project Manager, AI-Native Delivery
You are the single owner of a project's outcome: the client relationship, the scope, the schedule and the acceptance sign-off. You will use AI tooling to keep notes, specs and status honest, and you will be the person who says no when a change does not fit the scope.
Apply for this roleSenior Business Analyst, AI Solutions
You translate a business problem into an agentic workflow and a spec an engineer can build from — what the agent may do, what it must escalate, and how anyone would know it worked. You spend a lot of time in the client's actual process before you write a line of it down.
Apply for this roleFour steps, no homework marathon.
Apply
Email us the role you want and something you have built — repository, app, case study or portfolio. A CV is welcome but the work matters more.
Conversation
One call with the people you would work with. We go through your work, how you make decisions, and what the role actually involves day to day.
Practical exercise
A short, scoped exercise based on real work — a few hours, not a weekend. Use AI tools if you normally would; we will ask about your choices afterwards.
Offer
We discuss terms directly and give you a written offer. If it is a no, you get a straight answer and the reason.
None of these fit?
Write to us anyway. Tell us what you build and where you would fit — we read every speculative application.