Research Scientist (Member of Technical Staff)

Ironsite AI•San Francisco, CA

7h•$200,000 - $400,000•Onsite

About The Position

Ironsite is seeking a Research Scientist to join their team and contribute to building vision-language models (VLMs) that transform construction data into an intelligence layer. The role involves working alongside the Chief Science Officer and research team on training, benchmarking, and deploying state-of-the-art VLMs capable of interpreting complex construction sites. This position offers a unique research environment with a proprietary, large-scale, and expert-labeled dataset, a diverse compute footprint (edge, on-prem, cloud), and the opportunity to deploy models to production on active jobsites. The role is an early opportunity to shape the foundation of the company's AI development.

Requirements

2-4 years of hands-on research experience designing and training deep learning models, particularly transformer-based architectures. Industry experience, top-tier PhD program, or equivalent.
Deep expertise with modern deep learning frameworks (PyTorch, JAX, or similar) and strong proficiency in Python with solid software engineering fundamentals.
Experience working with large-scale vision or language datasets.
A track record of shipping meaningful research results, whether at a company, in a lab, or in publications.
A background in Computer Science, Machine Learning, AI, Robotics, or a related field.

Nice To Haves

Experience with vision-language models, video understanding, or multimodal architectures.
Hands-on experience with post-training techniques for large language or vision-language models (SFT, RL methods, parameter-efficient tuning such as LoRA).
Familiarity with the challenges of video data, including temporal reasoning and long-context modeling.
Publications at top-tier AI, ML, or CV conferences.

Responsibilities

Train and iterate on Ironsite's core models, designing, training, and iterating on vision-language models fine-tuned for spatial intelligence in construction environments.
Run experiments end to end, from data preparation through training, evaluation, and post-training, moving fast, measuring honestly, and sharing learnings.
Contribute to advancing the field by establishing new baselines, improving post-training recipes, or exploring long-context architectures.
Build and improve the Construction Intelligence Benchmark suite across video question answering, temporal reasoning, activity recognition, and site-level analytical reasoning.
Design evaluation metrics that measure real-world construction task performance.
Own the evaluation loop for personal work to ensure continuous improvement.
Ship models to production by taking research code to deployment, collaborating with infrastructure and hardware teams.
Apply distillation, quantization, and model routing techniques for efficient inference.
Learn from field data and production feedback to shape future experiments.
Grow with the team by learning from senior researchers, contributing to the team's intellectual culture, and developing a point of view on future research directions.
Read papers, share ideas, and help set the technical bar for the team.