What We're Building
Homeswift is rebuilding how Singapore - and eventually Southeast Asia - buys, sells, and rents property. We're not a listings portal. We're a data-intelligence platform: geospatial analysis, AI-driven search, real-time transaction data, and agentic workflows that do the hard work of home research on behalf of the user.
The frontend is the product. We're building the kind of interface that makes people stop and say how did they do that - interactive maps with live data overlays, streaming AI responses, complex filtering rendered at 60fps, and a conversational layer that feels like talking to someone who actually knows the Singapore property market.
We're a small team moving fast. You'll have real ownership, no bureaucracy, and direct line to the founders.
What You'll Work On
- Ask Swift - our AI property assistant. You'll own the streaming response UI, context-aware query handling, and the interface layer between the LLM pipeline and the user. This includes real-time token streaming, multi-turn conversation state, and tool-call rendering (when Swift runs a search or pulls a document, the UI needs to show that work happening live).
- Map Explorer - built on Mapbox GL JS. You'll push what's possible with WebGL-accelerated layers: transaction heat maps, Swift Score overlays, district-level analytics, property clustering at scale. Performance is a product requirement, not an afterthought.
- Agentic UI patterns - as we build out multi-agent workflows (parallel property search, document analysis, mortgage estimation), you'll design and implement the interface primitives that make agentic AI legible and trustworthy to non-technical users. This is genuinely unsolved design-engineering space.
- Data visualisation - property market data is dense. You'll build the charts, tables, and interactive components that turn URA transaction history, PSF trends, and profitability analytics into something a buyer can act on in 30 seconds.
- Performance - Singapore users are on mobile. Everything you ship needs to be fast on a mid-range Android on LTE. You'll own Core Web Vitals, bundle optimisation, and render performance across the stack.
Stack
Frontend
- Next.js 16+
- TypeScript - strict mode, no exceptions
- Tailwind CSS + Radix UI primitives
- Mapbox GL JS for geospatial rendering
- TanStack Query for client-side data fetching
- Jotai for client state
AI / Streaming
(to be confirmed - details shared during interviews)
- LLM API integration (streaming, multi-turn, tool use)
- Agentic pipeline integration
- Real-time AI response rendering
Infrastructure
(to be confirmed - details shared during interviews)
- AWS EKS, Aurora PostgreSQL, PostGIS
What We're Looking For
Must-have
- 1+ year building production React/Next.js applications - not tutorials, real users, real scale
- TypeScript fluency - you think in types, not around them
- Strong UI craft and interaction design instinct - you care about how something feels, not just how it looks. You've thought about states, transitions, feedback, and edge cases. AI can generate layouts it can't yet feel when something is off.
- Sharp eye for behavioural design - empty states, loading patterns, error handling, micro-interactions. The details that separate a product people trust from one they don't.
- Experience integrating LLM APIs (OpenAI, Anthropic, or equivalent) in a production product
- AI-native development workflow - you use AI tooling (Cursor, Claude, Copilot, or similar) as a force multiplier every day. Someone going manual in 2026 won't keep up here.
- Comfort with geospatial or mapping libraries (Mapbox or similar)
- Ability to read and reason about backend code - you won't be blocked waiting for someone else
Strong plus
- Experience building streaming AI UIs (token-by-token rendering, tool-call visualisation, optimistic updates)
- Familiarity with agentic workflows and how to make parallel AI work legible in a UI
- PostGIS or geospatial query experience
- Prior work at a PropTech, fintech, or data-heavy consumer product
- Singapore property market knowledge
What This Is Not
- A maintenance role. We're building new surface area constantly.
- A siloed role. You'll work directly with the founders, contribute to product decisions, and own features end-to-end.
- A role where AI features means a chat widget bolted onto an existing product. AI is load-bearing infrastructure here.
- A place for engineers who treat AI tooling as optional. We move fast because we use every available edge - if you're not already building with AI assistance as a core part of your workflow, this will not be a good fit.
Compensation & Benefits
- Equity - we're pre-Series A and the round will be structured to reflect early employees fairly
- Equipment budget - use what makes you fast
- Direct access to founders and no middle management
- The chance to own a meaningful piece of a product that people in Singapore will use to make one of the biggest decisions of their lives
