Machine learning and forecastsIndependent founder project
Marketplace Intelligence
A six-model pricing and demand system: recommends a price, estimates sale probability, and shows the evidence behind each decision.
Problem
Pricing and demand decisions need evidence.
Founder’s role
Dmitri trained the models and built the API and dashboard.
Result
Six models in one demo: pricing, sale probability and demand forecasts using public datasets.
Problem
Pricing work often lives in notebooks: one model for price, another for demand, no comparable listings, and no way for a non-ML user to see why a number appeared.
Solution
One FastAPI + React product where price, sale probability, demand forecast and comparable listings share one workflow; production champions are separated from weaker diagnostic modules.
Result
1.47M listings and 5M behavioural events: a team goes from scenario to price, comparables and model evidence without eight disconnected chat threads.
