Sai Dinesh Reddy
Maluchuru.

200 TB+ daily raw logs and 300 B+ single-day Blob transactions; pipelines that cut compute and storage by 70%.

Designed the LLM-as-a-Judge framework for video content — rubrics, scoring pipeline, prod rollout with ML + Applied Science teams.

Stood up the Data Engineering division from scratch — 200+ pipelines and ML workflows with monitoring + alerting.

Ran ODSP's biggest cost line — Highcharts dashboards surfaced $M-level savings and drove exec budget decisions.

Business-critical system used by 28+ Windows teams; drove Passkey adoption +30%.

Cut ML training costs 80% via Resilience LPG framework optimizations; experimentation days → hours.

Trained 50+ engineers on EDA with Claude and GPT-4; championed LLM adoption across ODSP.

Three foundational datasets — Node Processing, Package Attributes, Package Items — that became the source of truth for logistics analytics worldwide.

Root-cause attribution across millions of daily packages in India — improved promise accuracy, reduced customer-facing misses.

Instrumented + analysed the India pilot; dashboards drove the nationwide rollout decision.

Compliance and tax pipelines across regions with 99.9% accuracy and full auditability.

80+ downstream teams on Hoot / Redshift / Airflow / ECS; SLA adherence +25% on mission-critical AMZL datasets.

Led end-to-end migration of the batch data platform — 3× throughput, resolved data skew, cut infra cost.

Cross-Loans/Cards/Chase-Merchant P&L attribution model for leadership.

Interactive dashboards for Mortgage Banking, Cards, Merchant Services — 300+ daily users, -40% manual reporting.

Automated data-wrangling POC (-60% data-prep time); led team-wide ML fundamentals sessions.

ML for engine fault detection using multi-sensor parameter streams; text mining to establish reliable ground truth labels.