Ai Ml Engineer

Job description

**Role & responsibilities** **Job Description** High-Level Job Description 1) Essential experience: Forecasting + end-to-end model building - Proven delivery of **capacity/demand/time-series forecasting solutions** (e.g., workforce/headcount and capacity planning, volume forecasting, infrastructure utilisation, contact-centre demand). - Strong hands-on capability to **design, build, train, and validate models end-to-end** (not GenAI-only), including feature engineering, model training, tuning, and handover for deployment/productionisation. - Demonstrated depth in relevant forecasting techniques, such as **ARIMA/SARIMA**, **Prophet**, **XGBoost/LightGBM for time series**, **LSTM/Temporal CNN**, **hierarchical forecasting**, etc. (approach may vary; rigour and depth are key). 2) Strong machine learning fundamentals - Solid understanding of core ML concepts (e.g., **epochs**, **loss/error metrics**, **overfitting**, **cross-validation**). - Time-series best practices: **time-ordered train/test splits**, awareness of **data leakage**, and handling **trend/seasonality**, **missing data**, and **outliers**. 3) Forecast evaluation and operational readiness - Clear experience using forecasting metrics and validation practices, including **MAPE/SMAPE**, **MAE/RMSE**, **prediction intervals**, **backtesting**, and monitoring for **drift/retraining triggers**. - Ability to translate forecasts into operational decisions (e.g., **capacity planning assumptions**, **scenario modelling**, and **what-if driver analysis**). 4) Technical stack and engineering maturity - Proficiency with the Python forecasting/ML ecosystem: **pandas**, **numpy**, **scikit-learn**, **statsmodels**, **Prophet**, plus **PyTorch** or **TensorFlow** where deep learning is applied. - Production mindset: **version control**, **reproducibility**, **documentation**, and basic familiarity with **MLOps** practices (even if lightweight). **Mandatory Skills**ML**Desirable Skills**forecasting,Python,ML OPS,Pandas,Capacity Planning Thanks, Digvijay Sunil Digvijay.6.patil@coforge.com