AI / Full-Stack Engineer – Generative Design

UNLOCKLAND; Singapore; On-site; SGD 80,000 - 150,000; Full time

Job summary

UNLOCKLAND, a Singapore-based AI technology company, is expanding its AI and product engineering team in Singapore. We are building an AI operating system for real estate development and urban planning, combining geometry, regulatory data and LLM-powered workflows to generate viable development options.

The role focuses on core intelligence and generative systems, not a traditional full-stack position, with close collaboration with architects, urban planners and AI engineers to translate

Qualifications

• 3+ years of software engineering experience.
• Proficient in Python and TypeScript/JavaScript.
• Hands-on experience building products with LLMs.
• Experience integrating LLMs into real product workflows.
• Strong understanding of APIs, databases and modern web architectures.
• Comfortable designing algorithms and solving ambiguous technical problems.
• Able to work closely with product designers and domain experts.
• Strong product mindset and adaptability in a fast-moving startup environment.

Responsibilities

• Build core intelligence and generative systems behind UNLOCKLAND.
• Develop Generative Development Algorithms for site geometry, planning constraints and massing.
• Build LLM-powered user interactions and structured outputs.
• Collaborate with AI Product Designers and engineers to translate concepts into production software.
• Ensure reliability, scalability and maintainability of systems.
• Translate real-world development logic into scalable software solutions.

Skills

Python

TypeScript/JavaScript

LLMs

APIs & Databases

Algorithms

Product mindset

Startup environment

Job description

UNLOCKLAND is a Singapore-based AI technology company building an AI operating system for real estate development and urban planning.

Our platform helps real estate professionals analyse land, planning requirements, development potential and project feasibility — and generate development options across different real estate asset types.

We are part of the Harvard Innovation Labs ecosystem, with several key enterprise customers internationally. We are expanding our AI and product engineering team in Singapore.

The Role

About UNLOCKLAND

UNLOCKLAND is a Singapore-based AI technology company building an AI operating system for real estate development and urban planning.

Our platform helps real estate professionals analyse land, planning requirements, development potential and project feasibility — and generate development options across different real estate asset types.

We are part of the Harvard Innovation Labs ecosystem, with several key enterprise customers internationally. We are expanding our AI and product engineering team in Singapore.

The Role

We’re looking for an engineer to help build the core intelligence and generative systems behind UNLOCKLAND.

This is a not a traditional full-stack role.

A major part of your work will be developing systems that can understand a development site, interact with users through LLMs, interpret planning and design requirements, and generate viable development solutions across different asset types.

These may include:

Residential · Mixed-use · Office · Retail · Industrial · Hospitality · Master Planning

You’ll work closely with architects, urban planners, product designers and AI engineers to translate real-world development and design logic into scalable software systems.

The challenge is not simply to generate geometry.

It is to build systems that understand:

What the user wants → What can be built → What should be generated → Why the solution makes sense.

What You’ll Build1. Generative Development Algorithms

Design and implement algorithms that generate development solutions based on:

• Site geometry and constraints
• Planning and zoning requirements
• Setbacks and development controls
• Height and density constraints
• GFA / FAR / FSR requirements
• Building footprints and massing
• Building orientation and placement
• Circulation and access
• Unit / program mix
• Parking and amenities
• Asset-specific design requirements
• Commercial and development objectives

You’ll help build different generation strategies for different asset classes rather than relying on one generic algorithm.

2. LLM-Powered User Interaction

Build AI systems that allow users to communicate development intentions naturally.

For example, a user might ask:

“Create a residential development that maximises sellable area while maintaining good unit efficiency.”

or:

“Show me three mixed‑use development scenarios with different residential and retail ratios.”

Your job is to help build the system that can:

understand intent → structure requirements → identify missing information → interact with the user → call the appropriate tools/algorithms → generate options → explain the results.

This may involve:

• LLM orchestration
• Structured outputs
• Tool calling
• Agentic workflows
• Context management
• RAG / knowledge retrieval
• Planning and regulatory data
• Validation and guardrails
• Evaluation systems

3. Design → Engineering

Work closely with our AI Product Designers to turn ambitious product concepts into production-quality software.

Our designers may use Figma and vibe coding to rapidly prototype new experiences.

You will take those concepts and determine:

How should this actually work?

What architecture should we use?

What should be deterministic vs LLM-driven?

What needs a geometry engine or optimisation algorithm?

How do we make it reliable and scalable?

You should enjoy turning fast-moving prototypes into robust products.

4. Geometry & Spatial Intelligence

Depending on your background, you may work on:

• Computational geometry
• Geospatial analysis
• Parcel and site processing
• Building massing generationSpatial optimisation
• Constraint solving
• Parametric generation
• 2D/3D geometry
• GIS
• Map-based interfaces
• Design option generation and evaluation

Experience in architecture or computational design is helpful, but not required if you are a strong engineer who enjoys solving spatial problems.

What We’re Looking For

• 3+ years of software engineering experience
• Strong Python and/or TypeScript/JavaScript
• Experience building production software
• Hands‑on experience building products with LLMs
• Experience integrating LLMs into real product workflows rather than only building simple chatbots
• Strong understanding of APIs, databases and modern web architectures
• Comfortable designing algorithms and solving ambiguous technical problems
• Able to work closely with product designers and domain experts
• Strong product mindset
• Comfortable working in a fast-moving startup environment

Strong Advantages

Experience in any of the following would be particularly valuable:

Computational geometry

Generative design

GIS / geospatial systems

Optimisation / constraint solving

Architecture / AEC software

CAD / BIM

Three.js / WebGL

Mapbox / Cesium

Rhino / Grasshopper

Revit / Autodesk APIs

Agentic AI systems

RAG / knowledge systems

LLM evaluation and observability

You do not need experience in all of these.

We care more about whether you can understand complex problems and build working systems.

How We Work

We work closely across disciplines:

Urban Planner / Architect

defines how real development and planning workflows should work

AI Product Designer

turns those workflows into intuitive AI-native product experiences

AI / Full‑Stack Engineer

turns those concepts into reliable algorithms, AI systems and production software

You won’t just receive tickets.

You’ll be expected to understand the problem, challenge assumptions, propose technical approaches, prototype quickly and help shape the product.

Who This Role Is For

We’re looking for engineers who are excited by problems where there isn’t an obvious Stack Overflow answer.

For example:

Given an irregular parcel, planning constraints, target GFA, building typology and commercial objectives — how should an AI system generate and evaluate development options?

Or:

When should an LLM make a decision, when should it call a deterministic algorithm, and when should it ask the user for more information?

Or:

How do we turn an architect’s design logic into an algorithm that can generate thousands of viable development scenarios?

If these problems sound interesting, we’d love to talk.