Why your inspection model fails on the second shift
Lighting changes, operators change, and the camera gain follows. A short note on capture discipline and what to log so drift is diagnosable.
TODO: publish article
INSIGHTS
Engineering write-ups on edge deployment, industrial data and practical quantum. No thought leadership.
PLANNED WRITE-UPS
Cards are placeholders. Each becomes a full article once the underlying work is finished and reviewable.
Lighting changes, operators change, and the camera gain follows. A short note on capture discipline and what to log so drift is diagnosable.
TODO: publish article
Working backwards from conveyor speed and reject-gate position to a per-stage millisecond budget, and what that rules out.
TODO: publish article
How to construct a defensible classical baseline for a scheduling problem, so a quantum result means something afterwards.
TODO: publish article
What the term covers, which solvers belong in the category, and when they beat a general-purpose MILP solver.
TODO: publish article
Mounting, sampling rate, power and data paths for vibration monitoring on machines with no network and no spare I/O.
TODO: publish article
The handover checklist we use: weights, configs, schematics, runbooks and the retraining path. Written so the client can leave.
TODO: publish article
Editorial policy
Articles state the setup, the baseline and the measurement. Where a client cannot be named, the numbers are normalised and that is stated plainly. We do not publish predictions about the industry.
NEXT STEP
Tell us the problem. If we have solved it, we will write it up.