Profile Image

Hi, I'm Ram.

Staff / Principal Engineer | Agentic AI, LLM Apps & RAG | AI/ML | SRE & Observability for 120+ Apps (New Relic, Splunk, HoneyHive) | LangGraph · FastAPI · AWS · Kubernetes

Glad to see you here I'm based in Phoenix, Arizona. I'm a hands-on engineering leader and architect focused on agentic AI, with 18+ years of building and running mission-critical platforms in insurance, banking and the public sector.

Today I lead an agentic AI underwriting platform on AWS Bedrock (LangGraph, FastAPI, MongoDB Atlas Vector Search) that cut policy processing time by 60%. I also built the evaluation framework that gates every release: golden datasets, a Python test harness, deterministic checks, LLM-as-a-judge scoring and OpenTelemetry tracing in HoneyHive, all wired into CI.

I'm also the SRE point of contact for 120+ applications in Nationwide's Excess & Surplus / Specialty area, running a mandatory observability standard across New Relic and Splunk. Earlier I migrated 20 applications from on-prem to Kubernetes, architected the shared ESS authentication layer used by thousands of users daily, and led a 15-person onshore/offshore team.

Certifications: AWS Certified AI Practitioner · CKAD (Certified Kubernetes Application Developer) · SRE Foundation

Coming to personal life, I live with my wife and two kids. I love to spend time with my family and friends. My hobbies include playing pickle ball , watching movies, cooking , listening to music, reading books, and learning new technologies. And last but not least I love Hiking , sunsets and exploring new places.

My Skills

My work today sits where agentic AI, platform engineering and site reliability meet: building LLM applications and RAG pipelines, proving they work with rigorous evaluation, and keeping 120+ production applications observable and reliable.

AI / GenAI

LangGraph (multi-agent, stateful workflows), LangChain, RAG, Amazon Bedrock (Nova, Titan embeddings), MongoDB Atlas Vector Search, prompt engineering, AI document processing, GitHub Copilot

AI Evaluation & Observability

Evaluation framework design, golden datasets, deterministic test harnesses, LLM-as-a-judge (rubric scoring, judge calibration), human-in-the-loop review, hallucination & regression detection, HoneyHive, OpenTelemetry, online evals, token/cost & latency analysis

SRE & Observability

New Relic (APM, dashboards, alerting), Splunk, Prometheus, Grafana, P0 incident response, root-cause analysis, latency & performance tuning, Horizontal Pod Autoscaling

Cloud & Platform

AWS (EC2, CodePipeline), Kubernetes, Docker, Helm, Rancher, Harbor, Concourse CI/CD, Harness Next Gen, GitHub Actions, Jenkins

Backend & APIs

Python, FastAPI, Java 8/11/17, Spring Boot, Spring Security / OAuth2, microservices, REST & SOAP, Apigee, OpenAPI (OAS), Swagger

Messaging & Data

Apache Kafka, Apache Camel, IBM MQ, MongoDB Atlas, Oracle, DB2, MySQL

Frontend

Angular 10–17, TypeScript, JavaScript (ES6), HTML5, CSS3/SASS, Bootstrap

Practices

Solution architecture, system design, event-driven & microservices architecture, multi-agent system design, API-first design, observability architecture, Agile/Scrum, TDD, code review, SonarQube

Below is the list of some of my skills:

LG

LangGraph

RAG

RAG

AWS

Amazon Bedrock

API

FastAPI

HH

HoneyHive

NR

New Relic

SPL

Splunk

OTel

OpenTelemetry

Java

Python

Spring

Spring Boot

Oracle

REST-API

Angular

CSS

Bootstrap

APIGEE

Docker

Kubernetes

Concourse

Jenkins

Harness

Helm charts

Rancher

My Professional Experience

18+ years across insurance, banking and the public sector, from Java back ends to cloud-native platforms to production agentic AI.

Nationwide Mutual Insurance, Phoenix, Arizona

Nov 2024 - Present

AI Lead Developer & SRE Lead, Excess & Surplus / Specialty (ESS)

  • Own the target architecture for an agentic underwriting platform: event-driven FastAPI microservices on AWS, LangGraph for stateful workflows, MongoDB Atlas for transactional and vector data.
  • Designed the workflow as explicit agent state machines (ingestion → validation → risk scoring → quoting → approval routing), cutting manual policy processing time by 60%.
  • Built the RAG layer over underwriting guidelines and policy documents using Atlas Vector Search and Titan embeddings on Amazon Bedrock.
  • Designed the evaluation framework that gates every release: underwriter-labeled golden datasets, a Python test harness, deterministic checks, calibrated LLM-as-a-judge and human review, enforced as a CI gate.
  • Instrumented the platform with OpenTelemetry tracing in HoneyHive (prompt, output, tokens, latency, cost per step) with online evaluators and alerting.
  • Sole SRE point of contact for 120+ ESS applications: onboarding to New Relic and Splunk, coverage reporting to leadership, and latency/performance SME on P0 incidents (20+ major incidents resolved).
  • Python
  • FastAPI
  • LangGraph
  • LangChain
  • Amazon Bedrock
  • MongoDB Atlas Vector Search
  • AWS
  • HoneyHive
  • OpenTelemetry
  • New Relic
  • Splunk
  • Prometheus
  • Grafana

Nationwide Mutual Insurance, Phoenix, Arizona

Mar 2020 - Nov 2024

Technical Lead / Application Architect, ESS Application Portfolio

  • Owned build, run and modernization for 40+ applications across Excess & Surplus, Management Liability, Specialty, Surety and Program lines.
  • Defined the reference Kubernetes deployment architecture (Helm, HPA, ingress, Concourse, Harness Next Gen, Rancher, Harbor) and migrated 20 applications off on-prem.
  • Architected the shared ESS identity and access layer (SSO, role-based authorization, Spring Security, Angular route guards) used by several thousand users daily.
  • Introduced API-first design with OpenAPI contracts exposed through Apigee; led Java 8 → 17 and Angular 10 → 17 upgrades.
  • Led a 15-person onshore/offshore team while staying hands-on; kept ~90% unit-test coverage and was an early adopter of GitHub Copilot.
  • Java 17
  • Spring Boot
  • Angular 17
  • Kubernetes
  • Docker
  • Helm
  • Harness Next Gen
  • Rancher
  • Apigee
  • Oracle

Wisconsin Department of Health Services, Madison, Wisconsin

Mar 2017 - Mar 2020

Senior Java Developer, CARES Eligibility Platform

  • CARES determines eligibility and benefits for Wisconsin’s major public assistance programs (FoodShare, Medicaid/BadgerCare Plus, Child Care, W-2) and powers the myAccess mobile app.
  • Built microservices from scratch for myAccess, letting residents apply for benefits and view benefit history.
  • Built event-driven integrations with Apache Kafka and Camel; secured services with Spring Security and OAuth2.
  • Java 8
  • Spring Boot
  • Kafka
  • Camel
  • Angular
  • Oracle
  • DB2
  • IBM MQ
  • Docker
  • Jenkins

American Family Insurance, Madison, Wisconsin

Apr 2016 - Mar 2017

Technology Analyst, Billing & Payments

  • Designed and built Spring Boot REST microservices for Billing Center and Payment Manager, connecting multiple billing systems and payment channels (EFT, card, IVR, Pay Now).
  • Java
  • Spring Boot
  • REST APIs
  • Angular
  • DB2

Nationwide Insurance

Dec 2011 - Apr 2016

Technology Analyst, Agent Gateway, Span of Control & Agency Transfers

  • Developed the Agent Gateway suite, giving agents one interface for customer, policy, quote, billing and claims data.
  • Built Span of Control (agent hierarchy feeding sales reporting and commissions) and the Agency Transfers app; provided on-call production support and release management.
  • Java
  • Spring
  • Oracle
  • Informatica
  • Linux

Bank of Nova Scotia (Scotiabank)

Jun 2007 - Dec 2011

Software Developer, Mobile Banking, Alerts & Commercial Cards

  • Built back-end services for Scotiabank’s then-new mobile banking channel (informational, transactional and payment services); wrote designs and led code reviews.
  • Java
  • Spring
  • Oracle
  • JavaScript

Contact Me

You may please connect with me on LinkedIn to discuss potential opportunities and collaborations! happy to discuss.