Stop losing money on warehouse imbalances? Building an ML alternative to "dumb" rule-based systems.

Hi everyone,

I’ve spent the last 5 years in Machine Learning, and lately, I’ve been obsessed with a specific logistics nightmare: The "East Coast vs. West Coast" stock gap.

Most SMEs I’ve talked to are still using "rule-based" logic to balance inventory.

If stock at Warehouse A < X, move Y from Warehouse B. It’s conventional, rigid, and usually leads to either expensive overstock or missed sales.

I’ve built an ML-driven optimisation engine that treats stock balancing as a dynamic problem rather than a static rule. It analyses velocity and demand to determine the exact transfers needed to maximise sales while slashing holding costs.

I have a working script/prototype ready, and I’m looking for some "war stories" from the community:

How do you currently handle inter-warehouse transfers? (Spreadsheets? Gut feeling? Rigid ERP rules?)
What’s the biggest "hidden cost" you've hit when stock is in the wrong place?

I’m looking to run some "no-risk" data simulations for a few pilot users to prove the ROI. If you're managing multiple hubs and want to see if ML can beat your current setup, let’s chat in the comments!