Artificial Intelligence Intern

airyhr is an AI-native HR platform for Southeast Asian SMEs. Our first product automates the recruiting funnel end to end — sourcing, screening, outreach, and interview scheduling — and converts every hire into a payroll record. We are incubated by an established Singapore HR services group with an existing SME client base, so your work ships to paying customers from day one. This is not a research internship and not a training programme: you will own features that real customers use in their week. WHY THIS ROLE EXISTS The product is live and working for its first users. Taking it from "works" to "scales" needs someone who has actually built with LLMs — who knows the difference between a demo that impresses and an agent that can be charged for. We are deliberately hiring at the intern level for this. In AI-native product work the useful experience is 12–18 months old, not ten years old, so we would rather find someone one year into building with these tools and give them the core product than hire for a title. You will report directly to the founder. Our CTO (former Ant Group technical lead) partners with you on architecture and is closely involved throughout. WHAT YOU'LL DO • Build agentic recruiting workflows — sourcing, screening, outreach, scheduling. You own features end to end, from schema to UI. • Work on the orchestration layer — tool calling, structured outputs, evals, guardrails, and token cost control: the machinery that makes agents reliable enough to charge for. • Integrate channels and services — job boards, WhatsApp/Telegram, Lark, calendars; later, payroll rails (CPF, IRAS). • Harden what you ship — audit logging, masking of sensitive data, CI/CD, and testing discipline. • Sit in on real customer calls. You will hear the HR manager describe the problem in her own words, not receive it as a ticket. WHAT WE'RE LOOKING FOR • About 1 year of hands-on experience building AI-native products — as an engineer, a product intern, a founder, or on your own. Internships, freelance work, and self-initiated projects all count equally; we care what you built, not where you sat. • You have shipped something with an LLM that real people used — not a demo, not a weekend tutorial project. A side project with actual users counts. A course assignment does not. • You can talk concretely about how it behaved in the wild: how you designed the prompts and context, what your eval set looked like, how it failed on real inputs, how you fixed it, and roughly what it cost to run. • Solid full-stack fundamentals — you can deliver a feature from database schema to working UI on your own. Python or Node.js on the backend; React or Vue 3 + TypeScript on the frontend. • Fast, visible delivery. You can clearly describe what you shipped end to end in the past three months, and show it. • You move forward with incomplete requirements — asking the right questions and making reasonable assumptions rather than waiting for a full spec. WHAT WE DON'T REQUIRE • A degree in progress or completed. Students on industrial attachment, recent graduates, and career switchers are all welcome. • HR industry experience. The team brings 20 years of domain depth and will teach you. • A polished CV. Send us the repo, the demo link, or the Loom instead. NICE TO HAVE • Experience with HR SaaS, payroll, tax/accounting, or ERP systems • Cross-border or multi-country systems and configuration-driven design • Experience integrating third-party APIs or fixed-format file specifications • Ideas on applying models to data reconciliation and anomaly detection • Working comfort in both English and Mandarin (many of our customers are China-based companies expanding into Southeast Asia) COMMITMENT • 3–6 months, full-time, on-site in Singapore. Our team ships in the same room in the early months; this is not a remote or part-time role. • Available to start within the next 4–6 weeks. • Monthly stipend of SGD 1,000–2,000, depending on experience and what you can already ship. • Strong performance converts to a full-time offer with meaningfully broader ownership as the team scales. TECH ENVIRONMENT AWS (Singapore) · Claude / Anthropic API · Node.js / Python · Lark ecosystem. Stack choices are open for discussion — including yours. HOW TO APPLY Apply through LinkedIn, and in your application include one link to something you built with an LLM, plus two or three sentences on how it broke in front of real users and what you did about it. That paragraph matters more to us than the rest of your CV combined.