Federico Rao
MTRM
Explore a product interface for AI and machine-learning fashion-trend analysis, research workflows, structured signals, and decision support.
System shape
A machine-learning pipeline for demand signals in fashion. Social text is classified by a custom model trained on a domain-specific demand lexicon, screenshots are processed by a fine-tuned object detector to identify garments, and the combined signals are scored and stored in a document database. Predictions are served through a Python API and the whole stack is containerised for deployment.
What can be inspected
The record is about method: how the labelling scheme was defined, how detection and language signals are combined into a score, and how the pipeline is structured for repeated ingestion. Trend predictions are probabilistic and unvalidated against commercial outcomes here, so no claim is made about accuracy, lead time, or resulting revenue for any brand.
Page focus
Inspect a fashion-trend research interface that structures machine-learning signals, exploration, and decision-support workflows.
What to inspect
Review visible product behavior, workflow decisions, interfaces, and technical constraints separately from outcomes requiring client evidence.
Measurement and limits
Evaluate Mtrm with observable checks relevant to this route. Record material changes when the underlying offer, system, evidence, or dependency changes. Technical eligibility or deployment alone does not guarantee rankings, traffic, enquiries, revenue, accessibility compliance, or operational improvement.