// Data · San Jose, CA

Suresh
Ravuri

Building high-throughput data pipelines, real-time streaming systems, and AIOps infrastructure that turns raw events into business clarity.

View Experience Get in Touch

pipeline_monitor.sh

Pub/Sub

──▶

Dataflow

──▶

BigQuery

● LIVE ● 3M+ rec/hr

Kafka

──▶

Spark

──▶

Snowflake

● LIVE ● lag <3min

pipeline health78 / 100 nodes OK

0 % query cost cut

0 M savings driven

0 M records/day QC

0 + stores powered

// 01 — experience

Where I've
Built Things

May 2025 – Present San Jose State University Research

Research Engineer — AIOps Copilot

Building an agentic AIOps copilot using LLMs & RAG to detect and auto-diagnose ETL and dashboard issues (freshness, latency, cost) via lineage and observability.
Designed SLA-driven remediation recommendation flows surfacing ranked fixes from pipeline lineage graphs.

Aug 2023 – May 2025 San Jose State University Teaching

Graduate Teaching Associate — AI & ML

Mentored 70+ students in Machine Learning under Prof. Ali Arsanjani (Google), leading sessions on end-to-end ML systems.
Covered data pipelines, model training, evaluation, and APIs while reinforcing software engineering best practices.

Sep 2021 – Jul 2023 Cognizant → Walmart GCP · BigQuery

Data Engineer

Engineered real-time Dataflow & Pub/Sub pipelines processing 3M+ POS transactions/hr, reducing inventory lag from 2 hrs to 45 min across 4,500+ stores.
Optimized BigQuery partitioning & clustering, cutting query costs 45% and dashboard latency from 18s to 4s.
Built Cloud Functions data-quality checks validating 80M+ records/day, reducing anomalies 65%.
Delivered Looker dashboards & ML models, driving $2.6M savings and reducing feature generation from 3 hrs to 18 min.

Jul 2020 – Aug 2021 Cognizant → Aston Martin Snowflake · Kafka

Data Engineer

Led a centralized Snowflake warehouse managing 1.8M+ records from 4 factories, optimizing storage and retrieval.
Designed ETL/ELT pipelines with Airflow, Spark, Kafka, reducing data lag from 12 to 3 min across 30+ sources.
Automated hourly updates for 1,200+ users via CI/CD on Kubernetes, eliminating manual workflows.
Cut incident response time by 40% using Grafana & CloudWatch monitoring dashboards.

// 02 — skills

Tech Stack

⚡ Streaming & Pipelines

KafkaPub/Sub DataflowSpark Debezium CDCAirflow DagsterAWS Glue

🗄️ Databases & Warehouses

BigQuerySnowflake RedshiftClickHouse MongoDBMySQL Hadoop

☁️ Cloud & DevOps

GCPAWS AzureKubernetes DockerTerraform Jenkins

💻 Languages

PythonSQL JavaGo JavaScriptC++ PySpark

📊 Visualization & Modeling

dbtLooker TableauPower BI GrafanaCloudWatch

🤖 AI / ML

PyTorchLLMs RAGResNet OpenAI APIGemini NumPy / Pandas

// 03 — projects

Things I've
Shipped

👁️

A-Eye: AI-Powered Diagnosis

ResNet-50 classifier with RETFound foundation model for multi-class retinal disorder detection at 97.4% accuracy. Integrated Gemini 2.5 Pro TTS for a multimodal voice AI agent enabling real-time text-to-speech diagnosis.

PyTorchResNet-50RETFound OpenAIFlaskGemini 2.5

VIEW PROJECT →

🛒

Retail Analytics Dashboard

Real-time ETL pipeline streaming 100K+ retail transactions/min with Kafka & Spark, achieving <60s end-to-end latency. KPIs modeled with dbt and visualized in Tableau to surface stockout risks and top-selling SKUs.

KafkaSparkPostgreSQL AirflowdbtTableauAWS

VIEW PROJECT →

🤖

AIOps Copilot

Agentic copilot using LLMs and RAG for automatic detection and diagnosis of ETL failures, dashboard freshness degradations, and cost anomalies — with SLA-driven remediation recommendations via data lineage.

LLMsRAGPython Lineage GraphObservability

VIEW ON GITHUB →

🔭

More coming soon…

Currently building. Check GitHub for the latest work.

GITHUB →