Software Engineer (AI Agents / Vibe Coding)

Company Overview

- Join a legacy EdTech firm that is actively reinventing itself. Silicon Workshop Limited (思進電腦) builds platforms used in primary schools across Hong Kong, and is expanding AI work so the next generation of digital natives can learn, grow, and thrive.

Job Highlights

- Position: Full-time

- Core Focus: AI agents, LLM/prompt experimentation, next-gen prototyping under tech uncertainty

- Location: Strictly Hong Kong-based (In-office / Hybrid options available)

- Education: Bachelor’s Degree or above in a STEM-related discipline (Postgraduate degrees welcome)

Background

Are you a builder who focuses on the "what" and the "why" rather than just the "how"? Silicon Workshop Limited is looking for a full-time Software Engineer to join our AI team. You will design, test, and evaluate AI agents, LLMs, and modern IDE-assisted workflows to resolve scientific and technological uncertainty in primary-school education platforms — not routine feature factories or LIVE BAU support.

Key Responsibilities

- AI experimentation: Research, benchmark, and experiment with LLMs, autonomous agents, and Human–AI collaboration for primary-school education tools; document tech uncertainty, methods, and results.

- Agentic workflows & prompts: Design, test, and refine system prompts and agent workflows; build eval harnesses; measure correctness, refusal quality, security, and hallucination risk on fixed test sets.

- Controlled prototyping: Use AI-assisted coding tools (Cursor, Windsurf, v0, or internal Silicon Workshop agents) to prototype and iterate research builds in staging / sandbox environments before any production consideration.

- Safe integration research: Study how AI-generated components can interoperate with existing codebases and data models under school isolation and audit constraints — experimental integration on non-production paths, not day-to-day LIVE ops.

- Human–AI interaction eval (secondary): Evaluate whether teacher- and student-facing flows remain clear and safe when agents assist; treat UI/UX as an evaluation surface for findings, not as cosmetic product polish.

Requirements

- Education: Holder of a Bachelor’s Degree or above in Computer Science, Information Technology, Software Engineering, Data Science, or a related STEM discipline (as required for RTH-TC research talent eligibility).

- Research experience: Demonstrated experience (academic or industry) in researching, benchmarking, or experimenting with LLMs, prompt optimization, autonomous agents, or Human–AI collaboration. Research papers, technical reports, or a thesis in an AI-related field highly valued.

- The "Builder" Mindset: Portfolio of functional prototypes or side projects built with AI-assisted development workflows, with evidence of evaluation — not only demos.

- Technical Literacy: Enough depth in JavaScript/TypeScript, React, or Python to debug, review, and verify AI outputs and reject unsafe or incorrect results.

- Critical Thinking: Strong hallucination-checking and code-review skills; fail-closed mindset for school data and child-related content.

- Communication: Fluent in English and Cantonese to work with local product, design, and research stakeholders.

Work Location & Eligibility

- Strict Location Requirement: This position is based permanently in Hong Kong; research duties are performed in Hong Kong.

- Candidates must be legally eligible to work and physically present in Hong Kong.

- No fully remote, offshore, or cross-border remote arrangements will be considered.

Not in scope for this engagement

- Sales, marketing, market research, or subscription rollout

- Routine LIVE production support / helpdesk / BAU feature grinding without tech uncertainty

- Fully remote or offshore work

Interested parties, please send a full resume with availability, current and expected salary to us. All personal data collected would be used for recruitment purpose only.