Spotable builds an AI platform that turns buildings into structured data, quantities and quotations in minutes. Contractors use our product to measure in 3D, extract geometry, map materials and generate accurate take offs without climbing a ladder. This requires computer vision, geometry, machine learning, LLM reasoning and real world performance.

We are looking for an AI Engineer who can take research ideas and turn them into production-grade features. Someone who understands that accuracy, latency and scale matter when contractors rely on AI in real commercial workflows.

What you will work on

You build and improve computer vision models for segmentation, depth estimation, façade and roof element detection

You design and optimize ML pipelines that can run efficiently across web, edge or cloud environments

You integrate CV and AI outputs with our CSG and geometry engines so that take offs remain robust and consistent

You work with LLMs for tasks like text extraction from drawings, safety and code compliance reasoning, bill of materials generation, supplier matching and construction workflow automation

You optimize inference speed and model footprint so real contractors can use the platform during field sales or quoting

You build internal data tooling for dataset curation, labeling, evaluation and continuous retraining

You debug failure cases from real project photos, drawings, PDFs and supplier catalogs

You collaborate with 3D engineers, integrations engineers and product teams to turn AI into measurable business impact

Who you probably are

You have experience taking ML or CV models to production, not just research notebooks

You have worked with LLMs or large sequence models in production or applied contexts, ideally with construction, CAD, quoting, materials or compliance data

You evaluate models using metrics that reflect the real world rather than just academic benchmarks

You can translate research papers and prototypes into reliable production features

You care about inference speed, memory footprint, quantization, caching and optimization techniques

You communicate trade offs clearly to product and engineering teams

You prefer impact over hype and accuracy over novelty

Tech you might touch

PyTorch, TensorRT or ONNX Runtime for training and inference

OpenCV and Open3D for processing and reconstruction

CUDA for GPU acceleration when needed

FastAPI or gRPC for inference serving

Docker for reproducibility

Weights and Biases for experiment tracking

S3 GCS or lakehouse-style storage for datasets

LLMs such as OpenAI, Claude, Llama, Mistral, or fine tuned domain models for structured reasoning and extraction

Vector DBs or retrieval pipelines for supplier catalogs, BOM reasoning and quoting logic

Not a checklist. If you are strong but use slightly different tools you are welcome.

Example projects you might own

Boosting segmentation accuracy for complex roof geometries like dormers, skylights or chimneys

Adding façade material classification that feeds directly into quantity and pricing logic

Using LLMs to convert 2D drawings, PDFs or catalog data into structured take offs and product mappings

Optimizing inference so segmentation and LLM reasoning run interactively in the field

Building data pipelines for depth or multi-view reconstruction

Evaluating CV and LLM models against real contractor tolerance for error

Cleaning supplier catalog data and building reasoning layers for product matching

Deploying inference endpoints across multiple countries with performance constraints

Why this role matters

AI is how Spotable sees the world

LLMs are how Spotable understands context

If both fail, measurement fails and quoting breaks

If they succeed contractors trust the system and adoption explodes

Your work directly impacts reliability, trust and scale across Europe and the US.

Apply

Send your CV or GitHub or LinkedIn to pj@spotable.com and tell us what AI systems you have taken to production.

If it looks like a fit we move fast.

‍