Wood Defect Product Owner

Neural Grader is building AI for automated wood defect detection and grading in real industrial sawmill environments.

We are looking for an interdisciplinary profile: someone with a strong forestry / wood processing / lumber grading background , and digital competences who can own the quality loop of our AI defect models.

Ideally, the person should have graduated from ENSTIB or a similar forestry / wood engineering / wood science school.

This is not a classic support role and not a general coordination role.

The core mission is to own and improve the full data → model → customer feedback loop for defect detection:

• investigate real customer failure cases
• understand whether the issue comes from data, annotation, model limitations, or production conditions
• improve evaluation datasets and validation logic
• work closely with annotation and ML teams
• help turn raw customer feedback into clear, actionable priorities
• shorten the loop from “model got it wrong” to “fix validated in production”

You should be comfortable discussing wood defects with customers, reviewing real production cases, and working with technical teams to improve model quality over time.

Ideal background

• Forestry, Wood Engineering, Wood Science, or Wood Processing
• strong practical understanding of wood defects and sawmill reality
• attention to detail and structured thinking
• ability to work with data, examples, quality issues, and technical workflows
• good communication skills in customer and internal discussions

This role sits at the intersection of wood expertise, AI model quality, annotation quality, and customer reality .