Job Description:
• Own end-to-end experiments on Boon’s foundation model, from architecture design through self-supervised pretraining, supervised fine-tuning, and production deployment
• Design and evaluate novel multi-stage vision architectures for construction drawing understanding
• Drive decisions on backbones, decoders, fusion strategies, loss functions, and training regimes
• Run rigorous experiments with baselines, ablations, and held-out evaluation on real construction drawings
• Pursue research directions to improve accuracy across construction trades and scopes
• Move models from experimental notebooks into the production inference pipeline
• Work hands-on with PyTorch, YOLO, SAM, DINO, and other modern computer-vision stacks
• Collaborate with ML engineers on deployment, quantization, and serving
• Debug failures on customer drawings and incorporate findings into subsequent training runs
• Collaborate with synthetic-data, annotation, and infrastructure teams on data and compute needs
• Partner with engineering leadership on the accuracy roadmap and strategic direction
• Write internal research reports and present findings, trade-offs, and recommendations to engineering leadership
• Help determine data acquisition and annotation priorities
• Define evaluation datasets and metrics
• Identify failure modes and design experiments to address them
Requirements:
• 3–7+ years in computer vision research
• Strong track record of published computer-vision research or trained production computer-vision models shipped at scale
• Hands-on expertise in multi-modal dense prediction, including segmentation, detection, or joint vision-language tasks
• Production experience with modern vision-transformer backbones such as SAM, DINOv2/v3, CLIP, SigLIP, or similar
• Strong PyTorch fluency and experience training large vision models
• Ability to move models from research into production inference pipelines
• Strong deep-learning fundamentals, including optimization, loss design, regularization, and self-supervised learning
• Fluency in English, written and verbal
• Ability to collaborate during California business hours
• Experience with Graph Neural Networks or relational reasoning architectures is a plus
• Experience with text spotting, OCR, or scene-text detection is a plus
• Experience with LoRA, adapters, or parameter-efficient fine-tuning is a plus
• Experience with self-supervised pretraining is a plus
• Experience with engineering or technical drawings, document understanding, or layout analysis is a plus
• Contributions to open-source computer-vision research are a plus
• Published papers at top venues are a plus
Benefits:
• Meaningful equity in an early-stage, well-funded startup
• Publication-friendly — we support our team publishing their work at top venues
• Autonomy to propose, defend, and run your own experiments, shipping the winners to real customers
• Collaboration with a world-class team from top tech companies and research institutions
• Work at the intersection of computer vision research and real-world industry impact, in a trillion-dollar market
• Potential additional compensation including discretionary bonuses and/or equity in the company