Senior Data Scientist / ML Engineer – Forecasting & RMS Engines (3-6 Years)

About RevOpt RevOpt is building an AI-native Revenue and Profit Optimization System for the global hospitality industry. Our platform helps hotels make better demand, pricing, inventory, and commercial decisions through practical forecasting, decision-support logic, explainability, and operational workflows. We are an early-stage product company with a strong hospitality and revenue-management founding team. We are building the core RMS platform for pilot hotels and are looking for a hands-on data scientist who wants to work on real, high-impact forecasting problems. The role We are hiring a Senior Data Scientist / ML Engineer – Forecasting & RMS Engines to strengthen the forecasting intelligence layer of RevOpt. This is a hands-on, onsite Chennai role. You will work directly with the founders, backend engineer, and frontend engineer to build demand-forecasting workflows using real hotel operating data, validate them rigorously, and convert forecast outputs into reliable pricing and decision-support inputs. This is not a pure GenAI/RAG/chatbot role and not a pure research role. We need someone who has actually worked on time-series, demand, sales, inventory, or comparable forecasting problems and can take models from data exploration to repeatable production workflows. What you will do • Build, evaluate, and improve daily demand-forecasting models for hotel rooms across dates, room types, booking segments, channels, lead-time bands, and other relevant dimensions. • Work with founders to convert hospitality signals—seasonality, day-of-week, holidays, events, competitor context, booking pace, cancellations, lead time, length of stay, and business rules—into model features. • Develop and compare baseline and advanced forecasting approaches such as seasonal naive, ETS, ARIMA/SARIMA/ARIMAX, gradient boosting/XGBoost, Prophet, or equivalent methods where appropriate. • Design robust backtesting and model-evaluation frameworks, including rolling validation, horizon-wise accuracy analysis, forecast bias, WAPE/MAPE/MAE, and error diagnostics. • Help define whether forecasts are safe enough to influence pricing through shadow-mode testing, monitoring, data-quality checks, guardrails, and human override workflows. • Implement forecast outputs and data-driven pricing-support logic that can be consumed by backend services, dashboards, and hotel-facing workflows. • Work with backend engineering to productionise data/model pipelines, APIs, scheduling, retraining, logging, and monitoring. • Document model assumptions, feature definitions, evaluation results, configurations, and decisions so the platform remains reproducible and maintainable. • Support gradual evolution from rule-based pricing guidance to more sophisticated optimization and decision logic as the product matures. What we need • 3–6 years of hands-on experience as a Data Scientist, Forecasting Engineer, ML Engineer, or comparable applied role. • Demonstrable experience in time-series forecasting, demand forecasting, sales forecasting, inventory forecasting, capacity forecasting, or an equivalent sequential prediction problem. • Strong Python, SQL, pandas, NumPy, scikit-learn, and data-analysis skills. • Practical knowledge of forecast evaluation: backtesting, MAPE/WAPE/MAE, bias, horizon-wise performance, and error analysis. • Experience handling messy, incomplete, real-world operational data. • Ability to explain model choices, accuracy, risks, and recommendations clearly to non-technical business stakeholders. • Comfortable collaborating closely with engineering to build reliable, production-usable workflows. • Willingness to work full-time onsite in Chennai, six days per week, during the early product-building phase. Good to have • Experience with ARIMA/SARIMA/ARIMAX, ETS, Prophet, XGBoost or other gradient-boosting methods for forecasting. • Experience in pricing, revenue management, operations analytics, supply chain, retail, travel, mobility, fintech, or similar decision-heavy domains. • Exposure to MLflow, Airflow, Dagster, Docker, FastAPI, cloud data platforms, or model monitoring. • Experience working in an early-stage startup or product company. Data and IP discipline RevOpt works with sensitive business and hotel operating data. You must be comfortable following strict data-security and IP rules, including not sharing client data, proprietary algorithms, or confidential product logic with public AI tools or external systems. Why join • Be an early core member of a serious hospitality technology product. • Work on high-impact forecasting and commercial decision problems using real operating data. • Work directly with experienced revenue-management domain experts and a focused engineering team. • Build core product capabilities rather than isolated analytics reports. To help us assess genuine fit, please answer all five questions in your application or cover note. Generic responses, copy-pasted AI answers, or applications that do not address these questions will not be considered. 1. Real forecasting experience Describe the most important forecasting model you have personally built. Please state: • what was forecast (demand, sales, inventory, capacity, occupancy, etc.); • the forecast horizon and granularity (daily/weekly/monthly; product/location/segment, etc.); • the data and key features used; • the models you tested and why you selected the final approach; • how you evaluated it, including specific metrics and results. 2. Forecast validation Assume you are forecasting daily room demand for the next 30 days for three hotels. What checks would you complete before allowing the forecast to influence hotel pricing recommendations? Please cover validation/backtesting, accuracy metrics, bias, data-quality checks, monitoring, and how you would move from shadow mode to live use. 3. Production experience Describe one ML or forecasting model you helped move from notebook/analysis into a production or near-production workflow. What did you personally build, how was it run or deployed, and how did you monitor failures, data issues, or model performance? 4. Chennai and availability This is a full-time, onsite Chennai role, six days per week. Please confirm: • your current city; • whether you can work onsite in Chennai from joining; • your notice period; • your earliest possible joining date. 5. Compensation and authenticity check Please share your current CTC and expected CTC. Then, in one sentence and in your own words, name the three most important checks you would make before using a demand forecast to influence a pricing decision.