Title: Founding AI / ML Engineer
Location: Philadelphia, PA (Hybrid -3 days) - Remote
Employment Type: Full-Time
Salary: $180K + equity
About Company:
******** is building the AI-powered workflow layer for construction, one of the world's largest industries.
Headquartered in Philadelphia with a remote-first team and a newly opened Philadelphia hub, Company has recently closed a seed funding round (term sheets signed, announcement pending). The company has approximately $2M ARR on its core product and plans to grow revenue 10x over the next 12 months.
The team includes a Chief AI Officer, seven engineers (two AI/ML engineers and five full-stack engineers), and a founding team working directly alongside customers to build and refine the platform.
Construction has resisted previous attempts at automation for more than 15 years. Subcontractors the company's core customers have seen little meaningful software innovation during that time. Agents and multimodal AI present the first significant opportunity to transform the industry. Because construction diagrams and floor plans are not well represented in the training data of existing foundation models, team is making a core technical investment in training its own models.
About the Role:
- Guthrie AI is hiring a Founding AI / ML Engineer with deep expertise in Computer Vision to lead model development for the platform.
- Construction diagrams and floor plans are not included in the training data of off-the-shelf foundation models, making in-house model development a critical part of the company's technical strategy. This role will own that initiative.
- Approximately half of the role focuses on computer vision, including segmentation models (SAM-level work, ideally with deep 2D segmentation expertise), classifiers, and object detection for floor plans and technical drawings. The emphasis is on architectural improvements and model development rather than simple fine-tuning.
- The remaining portion of the role focuses on agentic orchestration and retrieval, including LLMs, fine-tuning small language models, RAG, and evaluating local versus foundation model trade-offs.
- This position reports directly to the Chief AI Officer. The existing AI/ML engineers (2–3) will report to this individual, making technical leadership and onsite collaboration highly valuable.
Key Responsibilities:
- Train and improve in-house computer vision models for construction diagrams, including segmentation, object detection, and classifiers.
- Lead model architecture development and improvements where off-the-shelf foundation models reach their limitations.
- Own the AI/ML model development roadmap and establish the technical direction for the AI/ML team.
- Build agentic orchestration and retrieval workflows, including RAG pipelines, fine-tuning, local versus foundation model selection, and cost optimization.
- Provide technical leadership and guidance to the existing AI/ML engineering team.
Requirements :
Must-Have:
- Deep expertise in computer vision with demonstrated experience in 2D segmentation. Experience with foundation models such as SAM, SAM3, or comparable models, with the ability to make architectural improvements beyond fine-tuning.
- Strong proficiency in Python, PyTorch, and CUDA, with experience shipping production-grade ML systems or a rigorous publication record in computer vision research.
- Proven ability to train and improve in-house models from scratch when off-the-shelf models are insufficient.
- Strong research background, with a publication record in computer vision preferred. Both research-focused candidates who remain hands-on and applied production computer vision engineers are encouraged to apply.
- Willingness to work onsite at the Philadelphia hub three days per week (flexibility available for exceptional candidates).
Nice-to-Have:
- Experience working with construction diagrams, floor plans, technical drawings, or architectural computer vision.
- Experience with agentic orchestration and retrieval, including LLMs, RAG, fine-tuning, small language models, and local versus foundation model trade-offs.
- Experience with object detection and classical classifiers.
- Published research in raster-to-sequence models, floor plan localization, structural priors, or related computer vision subfields.
- Previous technical leadership experience, including mentoring or managing AI/ML engineers.
