SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Higharc is a VC-backed startup transforming how new homes are designed and built. The company has raised over $175M from top-tier venture capital and 18+ strategic investors in construction, building products, and distribution. The founding team has shipped products for Autodesk, Electronic Arts, Nike, and Apple.
You will join the Platform R&D team as an Applied Research Engineer, responsible for translating spatial AI research into production capabilities used by homebuyers and builders. Your core mission is taking models proven in research and shipping them into the product.
Key responsibilities include:
- Build the component layer around the layout synthesis engine: create a documented API contract, a service the product can call, and an evaluation gate that runs on every change.
- Operationalize model retraining into a one-command job with built-in benchmarks and team-readable results.
- Own the latency and cost budgets for learned capabilities.
- Extend core research into new capabilities such as multi-story plans and real-time editing.
- Partner with infrastructure engineers on MLOps and deployment; mentor researchers and interns on engineering practices.
- Document and communicate work through weekly status updates, clear pull requests, and design notes before major changes.
You thrive when models you've worked on run in production. You excel at taking well-defined problems with attached metrics and closing them. You improve existing codebases without unnecessary rewrites. You think about user experience (e.g., what a 45-second wait means) before optimizing benchmarks. You're honest about evaluation, catching test-set artifacts and reporting regressions transparently. You communicate clearly in writing across research, product, and infrastructure conversations. You have domain intuition about buildings—knowing that a bedroom door opening onto a kitchen is wrong before metrics confirm it.
The company is entirely remote since 2018, with regular in-person team meet-ups across US hubs. Most collaborators are remote, some part-time, so asynchronous communication is essential.
REQUIREMENTS:
- 5+ years of professional software development, including at least 2 years training and shipping ML models to real users
- End-to-end model ownership: data preparation, training, evaluation, serving, and user-facing integration
- Fluency in Python and PyTorch (TensorFlow or Keras acceptable if willing to transition to PyTorch)
- Experience with geometric or structured data (floor plans, CAD, BIM, graphs, 2D/3D geometry)
- Built or maintained an evaluation harness or benchmark used to gate changes
- Packaged a model behind an API (FastAPI or similar) and owned its latency and cost
- Master's degree in computer science, machine learning, or related field, or equivalent shipped work (PhD not required)
- Comfort working remotely with a team on US Eastern hours
MAJOR PLUS:
- Background in architecture or building (degree, professional practice, or AEC software work)
- Experience with constraint solvers (OR-Tools CP-SAT), optimization, or generative models for layouts/floor plans
- Plugin/tooling work in Rhino, Grasshopper, or Revit; experience with Hugging Face, experiment tracking, cloud model serving (AWS or Modal); publications or open-source contributions
