Suresh

About Me

I'm a Solutions Architect & Senior AI Engineer specializing in Generative AI — shipping production RAG pipelines, multi-agent workflows, and agentic systems using LangChain, LangGraph, AWS Bedrock (Claude, Titan), and CrewAI.

I design cloud-native systems on AWS — ECS Fargate workloads, serverless APIs, data platforms with PySpark & Athena, and full IaC with CDK & Terraform. I operate on an Agentic Driven Lifecycle (ADLC) philosophy — agent-first design, automated testing, and guardrail-enforced AI governance — delivering solutions that are functional, secure, and production-proven with a 30% user engagement boost through AI-powered features shipped to production.

Connect on LinkedIn

Agentic AI

LangChain LangGraph AWS Strands CrewAI AutoGen n8n RAG (FAISS / Pinecone) MCP LangSmith Prompt Engineering

ADLC

Agent-first Architecture Multi-agent Orchestration Human-in-the-loop AI Guardrails Agentic CI/CD Responsible AI Cost Governance

Cloud & AI Platforms

AWS Bedrock ECS Fargate Lambda SageMaker Athena / Redshift CDK (TypeScript) IAM / STS cdk-nag

Languages

Python PySpark TypeScript Go Java

Backend / Frontend

Next.js React NestJS Spring Boot GraphQL REST APIs

DevOps & Data

Terraform Docker Kubernetes GitLab CI/CD Apache Spark PostgreSQL DocumentDB

Projects

Personal Projects

⚡

Serverless REST API & Lambda Power Tuning

Production-grade REST API on AWS — API Gateway, Python Lambda, DynamoDB — with Lambda Power Tuning to find the optimal memory configuration across 6 configs (128 MB – 3 GB).

⚡ 66× speedup — 112.7 ms → 1.7 ms by moving from 128 MB to 512 MB
💰 12× cheaper — $0.24 → $0.02 per million invocations at optimal config
🔬 Benchmarked across 6 memory configs using Lambda Power Tuning (Step Functions)
💡 Lambda memory controls CPU allocation — more memory unlocked near-instant Boto3 connections
🏗️ Two decoupled CDK stacks — runtime API and tuning pipeline independently deployable

CDK Lambda DynamoDB Step Functions Python

GitHub LinkedIn Post

Professional Work

🤖

Multi-Agent AI — Natural Language to Analytics

Production multi-agent system using LangChain, LangGraph, and AWS Bedrock (Claude/Titan) that translates natural-language questions into Athena SQL over enterprise S3 data lakes — enabling self-service analytics for leadership.

⏱️ Cut ad-hoc reporting turnaround from days to minutes for leadership teams
🤖 LangChain + LangGraph multi-agent orchestration with AWS Bedrock (Claude & Titan)
🗣️ Natural-language → Athena SQL translation over enterprise S3 data lakes
📊 Unlocked self-service analytics without BI team involvement

LangChain LangGraph AWS Bedrock Athena S3 Python

🧠

RAG Pipelines on ECS Fargate

Built and deployed production RAG pipelines with FAISS vector stores and Bedrock embeddings into an ECS Fargate microservices platform — secured for enterprise-grade AI delivery.

🔍 FAISS vector stores + Bedrock embeddings powering production semantic search
🔒 Enterprise security: IAM-based access control, VPC isolation, WAF v2 (OWASP)
🤫 Zero credential exposure via Secrets Manager integration
📦 Scalable ECS Fargate microservices architecture

RAG FAISS AWS Bedrock ECS Fargate WAF v2 Python

🔐

Keyless Multi-Account AWS CI/CD Platform

Architected a keyless multi-account CI/CD platform using CDK (TypeScript) and GitLab CI — centralizing IAM governance across dev/test/prod and eliminating long-lived credentials entirely.

🔑 Zero long-lived credentials via EC2 instance profile → STS AssumeRole with tagged sessions
🏛️ Centralized IAM governance with reusable policy factories + mandatory permission boundaries
🛡️ cdk-nag security validation enforced on every build
📋 CloudFormation naming guardrails enforced at execution layer

CDK GitLab CI IAM STS cdk-nag TypeScript

🚚

SQL Server to Cloud Data Pipeline

Architected PySpark-based data pipelines migrating SQL Server datasets to S3 Parquet + Athena + Glue, with automated nightly full-replace sync — eliminating manual reconciliation entirely.

🌙 Automated nightly sync via EventBridge Scheduler — zero manual reconciliation
🗄️ SQL Server → S3 Parquet + Athena + Glue full modernization
⚡ PySpark + JDBC + boto3 pipeline stack for scalable data processing
📈 Unlocked self-service BI across business teams

PySpark Athena Glue EventBridge S3 Python

🏗️

Terraform IaC — Zero Config Drift Platform

Provisioned end-to-end AWS infrastructure with Terraform IaC across dev/test/prod workspaces — covering 10+ services with DynamoDB state locking ensuring zero config drift at scale.

✅ Zero config drift across all environments via DynamoDB state locking
☁️ Full coverage: ECS Fargate, RDS, Cognito, Lambda, EventBridge, WAF v2, Glue, Athena, CloudWatch
🔀 Complete dev/test/prod workspace isolation with reusable modules

Terraform ECS Fargate RDS Cognito WAF v2 DynamoDB

⚙️

Zero-Downtime CI/CD with AI-Assisted Delivery

Delivered zero-downtime CI/CD via GitLab pipelines automating Docker builds, ECR pushes, and ECS task definition revisions on every commit — with AI-assisted prototyping accelerating delivery cycles.

🟢 Zero-downtime deployments on every commit via rolling ECS task revisions
🐳 Docker → ECR → ECS task definition revision fully automated pipeline
🤖 Claude Code for AI-assisted agentic workflow prototyping
🔐 Least-privilege IAM + Secrets Manager enforced throughout

GitLab CI Docker ECR ECS IAM Claude Code