Equip BILL's platform with the best-in-class document understanding capabilities leveraging state-of-the-art AI technologies to maximize value to SMBs.

Helped prototype and develop an MVP for an early-stage startup in the Supply Chain sector. Main contributions were in developing an enterprise-grade demand forecasting solution coupled with a RAG application to provide business value to supply chain planning teams.

● Managed data science & engineering teams for the Supply Chain AI SaaS product for 5 CPG customers. Forecasted unconstrained demand employing multiple forecasters (statistical methods and gradient boosted machines) and meta-ensembling techniques.

● Partnered with product and customer success leadership to establish key performance indicators, resulting in a 15% improvement of trial subscription conversions.

● Designed and implemented the end-to-end deployment playbook for the demand forecasting module, streamlining processes, and reducing deployment time by 30%.

● Collaborated with cross-functional teams to identify and address critical business challenges, leading to a reduction in forecasting errors by 10-15% through implementation of automated model refreshes and performance evaluation metrics.

● Provided mentorship to data scientists on my team, fostering their professional development and facilitating their advancement within the company.

● Led a team of data scientists and engineers to develop the Asset Health AI product which drove substantial asset maintenance cost reductions of ~20% at two steel plants. Implemented an advanced anomaly detection algorithm (multi-scale convolutional recurrent encoder-decoder networks) coupled with TICC segmentation.

● Spearheaded initiatives to standardise model training workflow (including onboarding the ML pipeline onto Kubeflow) improving developer efficiency by 4x.

● Interfaced directly with customers, including executive sponsors and plant managers, to deliver compelling results and articulate technical details with clarity at all levels.

● Lead data scientist and product owner of an energy prediction application designed for electric arc furnaces (EAF) at steel plants.

● Developed random forest models to accurately forecast energy consumption across multiple sections of steel plants and innovated an optimal trading strategy that achieved up to 10% reduction in energy costs and contributed to a more sustainable carbon footprint (US patent issued).