01Origin/ 10
Cartoon portrait of Yashas Suresh
control planeus-east-1
cloud onlineus-east-1 · College Station, TX

Hi, I'm Yashas

Cloud · DevOps · AI

Cloud / DevOps / AI Engineer

Kubernetes, Grafana, CI/CD — then RAG, LLMs, and agents on the same cloud rails.

KubernetesGrafanaRAGLLMsDockerAWS

Delivery factory

Commit → Grafana

Tap a stage — the pipe advances. No endless scroll required.

stage 01 / 05

Push to main — the story starts

Every change lands in git. Reviews, branches, and a clean history so the pipeline always knows what it’s shipping.

Toolkit

DevOps first. RAG & LLMs next.

Click the icons — same stack I ship with: pipelines and clusters, then grounded retrieval and agents on top.

01 · DevOps & cloud

02 · RAG · LLMs · agents

Page · Ops

DevOps that ends on a dashboard — not a hope

Kubernetes, Grafana, pipelines, and AWS. Cloud is the substrate; DevOps is how it stays alive after the demo ends.

k8s

Kubernetes platforms

Workloads that roll out cleanly — health checks, Helm, and clusters that forgive bad deploys with fast rollbacks.

cicd

CI/CD that actually ships

GitHub Actions, Jenkins, Docker images — every commit closer to a deploy-ready artifact, not a weekend surprise.

o11y

Grafana on the glass

Prometheus metrics and Grafana boards so on-call sees the truth before users feel the pain.

Page · RAG / LLM

Trained on retrieval. Shipped as agents.

Not just prompt demos — production RAG, embeddings, tool-calling LLMs, and Bedrock on the same AWS fabric as the DevOps stack.

  • RAG over AgriLife PDFs with pgvector — answers cite sources, not vibes
  • LLM agents that route tools: geo queries, XGBoost inference, retrieval
  • Hybrid search in one Postgres: PostGIS proximity + semantic embeddings
  • Bedrock / Nova pipelines with Textract OCR behind Terraform
  • Streaming FastAPI chat APIs consumed by a live Next.js geo client

Page · Catalog

Three planes. One cloud habit.

/k8s

Kubernetes & platforms

Container platforms, Helm, and cloud-native deploys — workloads that roll forward and roll back cleanly.

/cicd

CI/CD & observability

GitHub Actions, Jenkins, Docker, Prometheus, Grafana — pipelines that ship and boards that catch trouble early.

/cloud

Cloud + AI systems

AWS, Terraform, and grounded LLM agents sitting on the same rails as production — not orphan demos.

Page · About

From contact centers to field-grounded AI

MS MIS at Texas A&M. Previously Genesys & AWS at Carelon — Impact Award for keeping production alive under pressure. Now building geo-RAG research platforms on real cloud infrastructure.

Graduate Research Assistant

Mar 2026 — Present

Texas A&M AgriLife Research

  • Ingested ~20M groundwater rows into AWS RDS PostgreSQL via EC2 and ArcGIS APIs; Glue ETL + IAM-scoped access for research teams.
  • Hybrid retrieval in one Postgres: PostGIS proximity + pgvector semantic search over AgriLife PDFs.
  • FastAPI LLM agent routing to geo tools, XGBoost, and RAG — Dockerized on EC2 with CI/CD and a Next.js geo client.

Full Stack Developer (GA)

Dec 2025 — Present

The Bush School, Texas A&M

  • Maintain WordPress properties on WP Engine with WCAG accessibility and university brand standards.

Genesys & AWS Cloud Developer

Feb 2023 — Aug 2025

Carelon Global Solutions

  • Migrated a 500-agent contact center to Genesys Cloud on AWS with zero downtime.
  • Built IVR and routing on Lambda + API Gateway with metrics, logs, and traces for on-call.
  • Automated brittle third-party workflows — 56% less manual ops; Impact Award for a P1 cutover that beat SLA by 50%.

Page · Connect

Spin up a conversation

Roles, platforms, AI on AWS — write me at yashassuresh@tamu.edu.

endpoint ready

Yashas Suresh

Cloud / DevOps / AI Engineer

Phone
979-344-1237
Region
us-east-1 · College Station, TX