Senior Computer Vision Research Engineer

About the role

This position places you at the center of an AI lab dedicated to solving high-stakes problems in the construction sector by building visual intelligence systems. You will own the development of advanced computer vision models that automate the labor-intensive process of construction takeoffs, a task that currently relies heavily on manual human effort. The role grants you significant autonomy to design novel solutions, iterate rapidly, and see ideas move from prototype to production-grade systems. You will operate at a fast pace, making decisions that directly impact the accuracy and efficiency of critical project workflows. Your work will involve tackling complex challenges related to spatial reasoning and document understanding within detailed construction drawings. You will be responsible for ensuring that these systems are not only accurate but also reliable and scalable in demanding environments. This is an opportunity to shape the technical foundation of a new generation of tools for the industry.

Key facts

What you'll do

• Architect and implement state-of-the-art computer vision models specifically for detection, segmentation, optical character recognition, and high-level reasoning across intricate construction drawings.
• End-to-end management of the machine learning lifecycle, encompassing data acquisition, meticulous data labeling strategies, large-scale model training, rigorous evaluation, and optimized low-latency inference deployed in production environments.
• Drive systematic improvements in model performance by focusing on core metrics of accuracy, reliability, and inference speed, adapting to the vast diversity of drawing formats found in the wild.
• Partner closely with full-stack engineers and product managers to ensure seamless integration of computer vision models into customer-facing features that are used daily in real-world scenarios.
• Lead the investigation into model failures by mining real customer data, performing deep root cause analysis, and engineering targeted fixes that prevent recurrence.
• Rapidly incorporate cutting-edge research findings from the field and translate them into practical solutions, implementing code quickly and rigorously measuring the tangible impact of each change.
• Define the long-term architectural strategy for all ML systems, making key decisions that guarantee scalability, maintainability, and peak performance across various construction trades and use cases.
• Act as the technical translator between deep domain knowledge of construction trade workflows and the design of effective model architectures, guiding complex projects from conception to delivery.
• Provide hands-on technical leadership and mentorship to junior engineers, instilling best practices for robust system design, comprehensive testing, and meticulous code quality.
• Collaborate with product and engineering leadership to influence the technical roadmap, identifying opportunities and leading high-impact, cross-functional initiatives that move the needle.
• Serve as the final gatekeeper for model quality and readiness, enforcing strict standards that ensure only production-worthy models are released for critical customer bidding processes.

Requirements

• Demonstrate proven, hands-on experience training and deploying computer vision and deep learning models within real-world production environments, not just in research settings.
• Possess practical, battle-tested experience with core computer vision tasks including semantic segmentation, object detection, document AI processing, and reasoning with vision-language models.
• Show a consistent ability to deliver machine learning solutions that meet strict service level agreements, balancing performance with reliability under pressure.
• Exhibit an eagerness to engage with imperfect, messy real-world data, including handling scanned documents, handwritten annotations, and heavily distorted construction drawings.
• Adopt a pragmatic approach to research where the priority is shipping functional code and achieving measurable business impact over pursuing theoretical perfection.
• Maintain a strong work ethic capable of thriving in the fast-paced, high-pressure environment of a startup where responsibilities are broad and demands are high.
• Display rapid learning capabilities, quickly mastering new neural architectures and efficiently solving unfamiliar technical problems with limited guidance.
• Commit to working fully on-site in the San Francisco Bay Area for this role, understanding that in-person collaboration is key to the team's success.

Nice to have

There are no preferred items listed in the source material beyond the core requirements and skills.

Practical notes

This is an on-site, full-time position in the San Francisco Bay Area. Compensation is competitive and tailored to individual experience and contributions, with performance-based rewards included in the total package.