Artificial Intelligence Researcher

About Spacial:

At Spacial, we are transforming the construction and design industry with an all-in-one platform that connects architects, engineers, builders, and homeowners. We help professionals streamline workflows, cut costs, and move projects from design to construction faster and more reliably.

Join us as we make the built environment smarter, faster, and more sustainable.

About the Role

We are looking for a strong applied AI researcher to design and ship the next generation of Generative AI systems- large language and vision-language models used to reason over and structure complex information, not to generate media or content- for the built environment. This is a hands-on role at the intersection of research and product: you will frame hard problems, prototype rapidly, and drive prototypes into production. The most interesting challenges here are not about text alone- they are about getting models to reason reliably over structured, relational, and spatial information and to produce outputs a downstream system can trust.

What You'll Do:

• Design and build applied GenAI systems that combine large language models, multimodal understanding, and agentic reasoning to interpret complex inputs and solve real problems for our users.
• Develop methods that reason over structured, relational, and spatial data- turning complex technical documents and drawings into reliable, machine-usable representations.
• Architect robust agentic systems: orchestration, tool use, context engineering, and the evaluation harnesses that make model behavior dependable enough to ship.
• Own experimentation and evaluation end-to-end- define the metrics, build the test harnesses, and push accuracy, robustness, and cost-efficiency toward production bars.
• Partner closely with engineering to move research prototypes into production-grade, high-impact features.
• Contribute to our applied research direction and keep us current with emerging techniques.

What You Bring:

• PhD in Computer Science, AI, Machine Learning, or a quantitative field (Mathematics, Physics, EE, or similar).
• 5+ years of applied AI/ML experience with a focus on Generative AI, LLMs, and deep learning.
• A demonstrable record of research impact- publications at top-tier venues (NeurIPS, ICML, ICLR, CVPR, ACL, or similar) and/or patents in AI/ML.
• Demonstrated ability to build systems that reason over structured, relational, or spatial/geometric data- not just natural language.
• A track record of shipping AI systems into production, with the judgment to close the gap between benchmark results and real-world reliability.
• Depth in agentic system design, retrieval-augmented generation, context and tool engineering, and rigorous evaluation.
• Expert-level Python and strong fluency with major deep-learning frameworks (e.g., PyTorch).
• Comfort working with large, messy, real-world datasets.
• Exceptional problem-solving, creative thinking, and analytical skills.

Nice to have:

• Experience with multimodal models for understanding visual input (vision and language).
• A record of turning documents, drawings, or technical specifications into structured data.
• Affinity for the construction, design, or broader built-environment domain.
• Familiarity with running and optimizing AI workloads in production.

Why Join Us?

• Define and ship AI-powered solutions where fundamental research meets real-world product impact
• Work on hard, differentiated problems with the potential for significant global impact in construction and design.
• A collaborative, fast-moving startup with direct access to experienced founders.
• Flexible hybrid work arrangement.

If you are a builder-researcher who wants to push GenAI into production on genuinely hard problems, we'd love to hear from you!