Pune, India
"Google for IDeaS" — Smart Query Resolution & Hotel Price Forecasting
May–Jul 2019
Revenue management runs on two things: getting the right price, and getting the right answer fast. I built tools for both — a smart-pricing engine and an internal search product nicknamed "Google for IDeaS."
I built "Google for IDeaS," a client-assisting product that classified incoming client queries and predicted the resolution steps needed to solve them — essentially a smart search layer that got users to answers faster than manually searching documentation. Alongside that, I created a smart pricing management module that forecast hotel room prices, directly supporting IDeaS's core revenue management business.
On the technical side, I implemented TensorFlow for text summarisation, feeding the output through Keras for downstream text classification — a two-stage pipeline that condensed long client queries into their essential meaning before categorising them. It was hands-on exposure to how deep learning tools get chained together to solve a real business problem, not just a single-model demo.
