Sabarish Ravichandran Mehala

Building and deploying production ML systems for demand forecasting and operational decision-making: 1. Designed and deployed scalable forecasting pipelines, improving accuracy by ~8% MAE and enabling more reliable planning 2. Built a hierarchical forecasting system across multiple aggregation levels, improving consistency between global and local predictions and supporting better operational decisions 3. Built end-to-end ML systems from data ingestion to deployment and monitoring, ensuring reproducibility and stability in production 4. Developed real-time monitoring dashboards (Grafana + Flux) to track model performance and detect drift against business KPIs

Developed AI systems to automate document workflows and improve decision-making reliability: 1. Built an OCR + LLM pipeline, reducing manual processing time by 65% and improving turnaround speed 2. Designed a multi-agent system (FastAPI, Pydantic) with fallback handling, increasing system reliability from 92% to 99.5% 3. Converted prototypes into production-ready services, improving scalability, logging, and fault tolerance

Built predictive models and decision systems for energy forecasting and optimization. 1. Developed time series models with weather data, improving forecast accuracy by 18% 2. Designed decision pipelines combining forecasts and consumption, enabling ~12% projected CO₂ reduction and cost savings