Smart Pricing ML Model Needed
Budget: ₹1,500 – ₹12,500 INR
Hiring Brief: ML Engineer for Hotel Dynamic Pricing (MVP)
Objective:
Build and train a machine learning model that recommends optimal room prices for a hotel, aiming to increase revenue (RevPAR/ADR) and reduce manual pricing work.
Result We Need:
* A rolling 90-day pricing calendar with a daily update of recommended prices (per room type).
* A simple daily output file or API we can review and approve.
* Clear, short reasons (e.g., demand up/down, event impact) for each recommendation.
What the Model Should Use:
* Past booking data (12 months).
* Current on-the-books snapshot (what’s already sold).
* External Local demand signals (events/holidays; near-term weather; others if helpful).
What You’ll Do:
* Design features, train a tabular ML model (learning price–demand relationships/elasticity), and validate with time-aware splits.
* Produce daily predictions and a ranked price recommendation for each future date.
* Package the solution so it runs automatically once a day and outputs a clean file (or lightweight endpoint).
* Keep it explainable and stable; we’ll approve prices manually in Phase 1.
Success Looks Like:
* Consistent daily output we can act on.
* Sensible, stable recommendations aligned with demand patterns.
* Early signs of revenue lift and time saved for the manager.
Timeline:
* Start ASAP; MVP live by October, 2025.
To Apply (keep it short):
* 2–4 sentences on your approach to training/validation and how you’d turn it into a daily output.
* 1 example of relevant ML work (pricing/forecasting/tabular).
* Your availability and rough quote.
Objective:
Build and train a machine learning model that recommends optimal room prices for a hotel, aiming to increase revenue (RevPAR/ADR) and reduce manual pricing work.
Result We Need:
* A rolling 90-day pricing calendar with a daily update of recommended prices (per room type).
* A simple daily output file or API we can review and approve.
* Clear, short reasons (e.g., demand up/down, event impact) for each recommendation.
What the Model Should Use:
* Past booking data (12 months).
* Current on-the-books snapshot (what’s already sold).
* External Local demand signals (events/holidays; near-term weather; others if helpful).
What You’ll Do:
* Design features, train a tabular ML model (learning price–demand relationships/elasticity), and validate with time-aware splits.
* Produce daily predictions and a ranked price recommendation for each future date.
* Package the solution so it runs automatically once a day and outputs a clean file (or lightweight endpoint).
* Keep it explainable and stable; we’ll approve prices manually in Phase 1.
Success Looks Like:
* Consistent daily output we can act on.
* Sensible, stable recommendations aligned with demand patterns.
* Early signs of revenue lift and time saved for the manager.
Timeline:
* Start ASAP; MVP live by October, 2025.
To Apply (keep it short):
* 2–4 sentences on your approach to training/validation and how you’d turn it into a daily output.
* 1 example of relevant ML work (pricing/forecasting/tabular).
* Your availability and rough quote.