Careers
Build something that has to be right
Deep Precision is building AI that reads joinery drawings and produces accurate cost estimates. Not chatbot-plausible. Actually right. We're hiring the team to make that happen.
Why join
Why Deep Precision
Greenfield, not maintenance
You're not inheriting a legacy codebase. You're designing the system from scratch. Every architectural decision is still on the table.
AI that has to be correct
This isn't summarisation or chat. Our AI reads technical drawings and produces cost estimates that real businesses rely on. Accuracy is the product.
Early-stage impact
Small team, big decisions. You'll shape the product, the architecture, and the direction. Not just write code someone else specced.
Open roles
Open roles
All roles are London-based. Remote flexibility for the right person.
Engineering
London · Full-timeTechnical Lead
What you'll do
- ●Own the AI estimation pipeline end-to-end, from PDF and DWG ingestion to structured Excel output
- ●Design and build the document understanding system that extracts materials, hardware, dimensions, and metadata from production drawings
- ●Make critical architecture decisions: model selection, parsing strategies, data pipelines, output formats
- ●Build evaluation and validation frameworks to measure estimation accuracy against real-world benchmarks
- ●Work directly with joinery and construction firms to understand their drawings and refine the system
- ●Set the technical direction for the company as we scale from MVP to production
What we're looking for
- ●Strong Python background with production system experience
- ●Experience with computer vision, document AI, or structured data extraction from unstructured sources
- ●Comfort with LLMs and prompt engineering, ideally in contexts where accuracy matters more than fluency
- ●Ability to work across the full stack when needed (we're a small team)
- ●Track record of building systems, not just models
- ●Bonus: any exposure to construction, manufacturing, or technical drawing interpretation
PythonComputer VisionLLMsDocument AIPDF / DWG ParsingData PipelinesopenpyxlAirtable
Engineering
London · Full-timeAI / ML Engineer
What you'll do
- ●Build and fine-tune vision models for technical drawing comprehension
- ●Design structured extraction pipelines that pull precise data from complex, multi-layered drawings
- ●Create evaluation frameworks and feedback loops so the system improves with every drawing it processes
- ●Develop the training data pipeline: labelling strategies, synthetic data generation, active learning
- ●Benchmark model performance against human estimators
- ●Research and prototype new approaches to document understanding
What we're looking for
- ●Strong ML fundamentals with hands-on PyTorch experience
- ●Experience with vision models (object detection, segmentation, or document understanding)
- ●Familiarity with RAG architectures and structured output from language models
- ●Interest in building evaluation-driven systems where you can measure whether the AI is actually right
- ●Comfort with messy real-world data (construction drawings are not ImageNet)
- ●Bonus: experience with MLOps, model deployment, or building ML systems that serve production traffic
PyTorchVision ModelsRAGStructured ExtractionEvaluationMLOpsHugging FaceONNX
Apply
Interested?
We don't do cover letters. Tell us what you've built and why this interests you.