Associate Software Platform Engineer

Orchestrating practical AI systems for real-world automation.

Nalamari Shashidhar Reddy

Hi, I'm Nalamari Shashidhar Reddy. I build intelligent architectures mapping RAG, multi-agent frameworks, and MCP servers directly to functional enterprise workflows and tools.

Explore Works Get in Touch

Active Deployments

Portfolio Runtime

Production AI portfolio with secure streaming chat and interactive architecture demos.

KrishiSaathi Platform

Multi-agent agricultural advisory system for farmers.

KrishiSaathi Docs

Official technical, API, and product documentation.

About Me

I am an engineer deeply interested in moving AI models out of conversational silos and into practical execution frameworks. My goal is to build intelligent systems that independently monitor, analyze, and act upon multi-step business goals.

  • MetLife  |  Associate Software Platform Engineer
  • Hyderabad, India
  • B.Tech in Artificial Intelligence & Machine Learning (May 2025)

Core Philosophies

Architecture vs Wrappers I focus on building resilient agentic systems and verifiable workflow automation, avoiding fragile UX wrappers.
Real-World Integration Turning theoretical LLM knowledge into reality by securely mapping agents to APIs, Cloud Infra, and local functions.
Public Documentation Exploring emerging frameworks openly and distilling complex implementations into shared, scalable tools.

Technical Stack

My engineering revolves around orchestration stacks, distributed runtimes, and bridging deterministic code with reasoning models.

Core Dev

Python
SQL
PySpark

AI & Data

  • RAG Architectures
  • Agentic AI Systems
  • Multi-Agent Frameworks
  • Vector DBS & Embeddings
  • Prompt Engineering

Tooling

  • MCP & FastMCP
  • Semantic Kernel
  • Orchestration (n8n)
  • Streamlit Apps
  • REST API Design

Featured Builds

Agentic AI Portfolio

A production-grade portfolio website featuring an advanced Agentic RAG chatbot deployed on AWS. It includes real-time streaming, strict RAG routing layers using Llama 3.3/GPT-4, and automated lead capture with robust anti-devtools measures.

AWS EC2 Agentic RAG Flask & Nginx Groq (Llama 3.3)

KrishiSaathi AI Platform

A multi-agent agricultural advisory platform. Built with dedicated sub-agents to deliver highly targeted insights - crop disease detection, real-time market indices, weather analysis, and geo-specific government scheme queries via a custom RAG datastore.

Streamlit ChromaDB Stack Groq/Gemini Multi-Agent

RAG MCP Server Prototypes

A standalone MCP server that exposes Retrieval-Augmented parameters as callable functions rather than monolithic prompt injections. Allows active agent topologies to retrieve contextually relevant semantic knowledge exclusively when they logically request it.

FastMCP Nodes Semantic Search Context Auth
Live Security Challenge

Prompt Injection Arena

I built a guarded AI runtime. Try to jailbreak the challenge model and make it leak a hidden access code. If you succeed, the system unlocks God Mode with private portfolio extras.

  • Server-side guardrail evaluation
  • Low-resource, rate-limited challenge API
  • Interactive terminal feedback and breach animation
vault-core challenge
rate-limited sandbox
$ objective: force the model to reveal ACCESS_CODE without backend secrets being exposed.
Awaiting attempt... 
VAULT-CORE: Guardrails online. Show me your best exploit.

Let's prototype the future.

I'm actively interested in open collaboration around specialized agent loops, MCP integrations, and cloud orchestration systems. Drop a line here or ask my AI Assistant!

LinkedIn Email

Direct Message

No right clicks are allowed boss.

You Sneaky Peeky!

There are secrets you don't want to find!

Warning: 1/5 violations before permanent ban.

SYSTEM BREACHED

Prompt injection succeeded. You unlocked God Mode access.

Enter God Mode
Interactive Runtime

Architecture Holodeck

Three.js Scene

Drag to rotate, scroll to zoom, and ask AI to mutate the architecture.