Entity Embeddings Deep Neural Network for Intermittent Demand Forecasting

In the context of the haute couture fashion industry 👗, the demand for products is often unpredictable and intermittent. Chances are that the products available today were not sold in previous years. This means, little to none availability of time-series 😨.

How could Machine Learning be effectively applied in this highly-challenging scenario 🤔 ?!

Enters Entity Embedding Deep Neural Network, a novel model which learns to reconstruct usefull time-series leveraging commercial similarities between current and past product, in an end-to-end way.

Want to know more? Here's the paper for all the math details 🤓.

ML Powered In-Store Real-Time Inventory using RFID

In this project, the goal was to evaluate the applicability of Machine Learning techniques to the real-time inventory management in public stores 🏬.

With item-level RFID tags, the challenge was to be able to quickly distinguish items available to customers from those in the back office. All while avoiding the expensive and time-consuming installation of physical shields. Less cost 💵, more flexibility 🏋, faster results 📈.

The project yielded very promising results, which led to the filling of a patent application.

Want to know more? Here's the paper for more details 🤓.

AI Image Tracking System in The Artisanal Food Industry

Research Question: is it possible to track each individual ham through the curing process?
Answer: yes! And with very interesting commercial implications!

Through the effective and efficient use of Transfer Learning techniques, the application of neural networks for real-time object detection, it has been shown that it is possible to recognize and track the individual ham, without creating disruptions or impairments to existing process.
Benefits 🦾: thanks to the new system, the final commercial quality of the ham can be predicted well in advance and monitored.
This leads to clear competitive advantage 💵.

📧 Contact me