Data Scientist Level 2 | 3-5 YOE | Immediate Joiners | Bangalore (Bengaluru)

About the Role Our client is seeking a Data Scientist (3-5 Years) to join their innovative team (FinOps Domain). This role involves owning the science layer of their AI-native Value-Control platform, which focuses on forecasting cloud costs, detecting anomalies, and powering optimization recommendations. The successful candidate will work closely with engineering teams to transition models from prototype to production, ensuring actionable insights for customers. Key Responsibilitie - sAnalyze cloud cost and usage data from the client subscribers, transforming raw data into model-ready inputs - .Design, train, and evaluate forecasting, anomaly detection, and optimization models - .Conduct rigorous experiments to validate model performance and savings estimates - .Package models for production, build scoring pipelines, and monitor for model drift and data quality issues - .Translate model outputs into transparent narratives for product teams and customers, ensuring trust and usability . Required Skil - ls3+ years of experience building and deploying ML models in production environment - s.Strong foundation in time-series forecasting and evaluation methodologie - s.Solid statistical knowledge, including hypothesis testing and uncertainty quantificatio - n.Proficiency in Python (e.g., scikit-learn, statsmodels, PyTorch) and SQL for data manipulatio - n.Experience with MLflow or similar tools for experiment tracking and model deploymen - t.Ability to work with complex billing or operational data to create clean feature pipeline - s.Excellent communication skills to convey model behavior and trade-offs to technical and non-technical stakeholder s. Preferred Ski - llsHands-on experience with Databricks, including job management and cluster configuratio - ns.Familiarity with cloud billing data from AWS, Azure, or G - CP.Experience in anomaly detection on operational or financial time-series da - ta.Knowledge of optimization techniques and operations research metho - ds.Und .