Sahith Reddy Vellenki

Owned production AI/ML risk-decisioning workflows across feature pipelines, model inference, evaluation, and monitoring. Enhanced Python-based feature pipelines for reliable model inputs in AWS environments, built and maintained inference/monitoring workflows, and led model evaluation by creating evaluation datasets and comparing candidate configurations for production rollout recommendations. Collaborated with Risk and ML engineering teams and applied production CI/CD practices to improve model readiness and scalability.

Supported AI/ML Trust & Safety workflows through AWS-based feature data pipelines, offline evaluation, and production inference monitoring. Built modular Python components for preparing and validating training data, contributed to monitoring and health checks that reduced manual review time by 20%, and prepared evaluation datasets to support candidate model comparisons ahead of production rollout.

Supported university research and academic operations using Python automation, data processing, web application development, testing, and data validation across student and faculty workflows. Built a tool to generate award letters and draft emails from spreadsheet data, reducing document processing time by 98%. Contributed to a web application for managing doctoral student workflows and developed data-processing pipelines to transform spreadsheet inputs into structured outputs, including rigorous testing and record verification for accuracy.