Looking for a job

Aravind Veluri

@aravindveluri

Message

Senior Member of Technical Staff at Oracle OCI Compute; scaled observability to 500K+ events/sec.

India

Message

At Oracle Corporation, I architected the Apache Flink layer of OCI’s Compute Observability Platform across 30+ production regions, scaling it to 500K+ events/sec and cutting AI infrastructure telemetry latency to a 3-minute P90. I also led a 4-engineer team to consolidate 9 components into a dependency-ordered pipeline, reducing per-region rollout time from 1 week to 4 hours.

I delivered a full-stack GPU fleet admin platform and a Java/Dropwizard query control plane, and redesigned permissions caching to reduce privileged-access propagation from 7 days to under 10 minutes. Earlier, at Amazon Inc., I supported a Spark reconciliation pipeline and implemented infrastructure-as-code migration; at Adobe Inc., I engineered an infographic reflow pipeline that led to a granted US patent.

Experience

Work history, roles, and key accomplishments

Current

Senior Member of Technical Staff

Current

Oracle Corporation

Jul 2023 - Present (3 years 4 months)

Architected the Apache Flink layer of OCI's Compute Observability Platform across 30+ production regions, scaling to 500K+ events/sec and cutting AI infrastructure telemetry latency from 15 minutes to a 3 minute P90. Led a 4-engineer team to consolidate 9 components into a dependency-ordered pipeline, reducing per-region rollout time from 1 week to 4 hours.

Software Development Engineering Intern

Amazon Inc.

Jan 2023 - Jun 2023 (5 months)

Implemented the migration from semi-automated provisioning to full IaC (CloudFormation, CDK) for a financial reconciliation platform processing over $80M/day in invoice and manifest records. Supported Parquet data ingestion in Apache Spark reconciliation pipeline by building a config-driven framework, onboarding 4 new data sources tracking $10M/day in invoices.

CloudFormation CDK Apache Spark Parquet Infrastructure as Code AWS Python Java

Research Intern

Adobe Inc.

May 2022 - Jul 2022 (2 months)

Engineered a deep-learning pipeline for infographic reflow across devices (YOLO-v5 detection, transformer-based reading-order model, GAN layout generation) on a 27,000 image synthetic dataset generated in C#/Python, which was granted a US patent.

Deep Learning YOLO V5 Transformers Gan C#Python Synthetic Data Generation