Artificial intelligence is rewriting every industry it touches - but construction, the second largest sector in the global economy, has barely felt it yet. That's not a problem. That's an opening. ZeroRFI is the AI company building the intelligence layer for the built environment - the platform that makes every building project smarter, faster, and more predictable than the last. We're looking for a Principal Software Engineer to architect the systems that put AI at the center of how buildings get designed, built, and operated. This isn't AI as a feature. It's AI as the foundation.
Key Responsibilities
- Design and implement generative AI models for automated building design, including floor plan generation, facade design, and structural optimization using state-of-the-art architectures (diffusion models, transformers, GANs).
- Develop computer vision pipelines for design and drawing analysis using modern frameworks like YOLO, SAM, and NeRF-based 3D reconstruction.
- Build graph neural networks and geometric deep learning models for structural analysis and MEP (Mechanical, Electrical, Plumbing) system optimization.
- Create reinforcement learning systems for multi-objective building optimization (energy efficiency, cost, occupant comfort, sustainability metrics).
- Integrate AI models with industry-standard BIM tools (Revit, Rhino/Grasshopper) through custom APIs and plugins.
- Deploy production ML pipelines using modern MLOps practices, including experiment tracking (Weights & Biases, MLflow), model versioning, and A/B testing frameworks.
- Implement physics-informed neural networks for building performance simulation and predictive modeling.
- Collaborate with architects and engineers to ensure AI systems produce practical, code-compliant, and constructible designs.
- Lead research initiatives and publish findings to establish us as a thought leader in AEC AI innovation.
Requirements
- Master's degree or PhD in Computer Science, AI/ML, Computational Design, or related field (or equivalent industry experience).
- 3-5+ years of hands-on experience building and deploying ML models in production environments.
- Deep expertise with modern deep learning frameworks (PyTorch preferred).
- Strong foundation in computer vision, 3D geometry processing, and spatial reasoning algorithms.
- Experience with generative AI models (VAEs, GANs, Diffusion Models, Transformers) and their practical applications.
- Proficiency in Python and scientific computing libraries (NumPy, SciPy, scikit-learn, Open3D, trimesh).
- Experience with cloud ML platforms (AWS SageMaker, Vertex AI, or Azure ML) and distributed training frameworks.
- Understanding of optimization techniques (genetic algorithms, gradient-based optimization, constraint satisfaction).
- Strong software engineering practices and experience with containerization (Docker) and orchestration (Kubernetes).
- Excellent communication skills to translate complex AI concepts to domain experts and stakeholders.
Preferred Qualifications (A plus, not a requirement)
- Experience with computational design tools (Grasshopper, Dynamo) and parametric modeling.
- Familiarity with building information modeling (BIM) standards and IFC data schemas.
- Knowledge of graph neural networks (PyTorch Geometric, DGL) for structural and spatial analysis.
- Experience with physics simulation engines (Mujoco, Isaac Sim) or FEA integration.
- Background in multi-agent reinforcement learning for complex system optimization.
- Contributions to open-source ML projects or published research in relevant venues (NeurIPS, ICML, CVPR, or domain-specific conferences).
- Experience with point cloud processing and 3D scene understanding (PointNet++, DGCNN).
- Understanding of construction workflows and building codes.
What We Offer
- Compensation that reflects the role: Base salary $250,000-$300,000 plus meaningful early-stage equity. You're not joining a company where the upside is already priced in - you're one of the people pricing it.
- Hybrid work, your way: San Francisco or Atlanta, with a one-time home office stipend to build the setup you actually want.
- Healthcare done right: Comprehensive coverage with multiple plan options - we're not making you figure it out yourself.
- Unlimited PT