Machine Learning Engineer, Vision

About the role

Foreman's detection models are trained in-house and run inside factories on the company's on-site software, watching live camera feeds around the clock. This role owns that model program: data strategy, training, evaluation, and release.

It sits at the center of the product. Every alert Foreman sends begins with a model this role is responsible for, and every capability on the roadmap begins with a model this role will build.

What you'll do

• Own the training pipeline for Foreman's vision models: data curation, labeling strategy, training, evaluation, and release.
• Expand detection coverage across the roadmap: fire and smoke, intrusion, safety compliance such as helmets and vests, and the capabilities that follow.
• Design evaluation gates on real factory footage that decide when a model ships.
• Optimize models to run reliably on standard on-site hardware without sacrificing accuracy.
• Study failure cases from production sites and turn them into measurable model improvements.

What we're looking for

• Experience training and shipping detection or vision models into production.
• A rigorous approach to evaluation: test sets, metrics, and regression discipline that hold up under real-world conditions.
• Comfort owning outcomes end to end, from raw footage to a model running on hardware in the field.
• Clear written communication about experiments, trade-offs, and results.
• Based in India. Remote, with occasional site visits.

How to apply

Email hello@foremanintelligence.com with "Machine Learning Engineer, Vision" in the subject line. Include your background, links to relevant work, and a short note on why this role. Every application receives a response.