// principal engineer

Atharv Chavan

Breaking AI models so you don't have to.

// what i do

Building at the intersection of security and AI

AI security & red teaming

Testing the boundaries of large language models. Finding the vulnerabilities before the bad actors do. Shipping fixes that actually hold.

LLM Security Guardrail Bypass Prompt Injection MCP Security Threat Modeling

Full stack engineering

From React frontends to Flask APIs. Whatever ships the thing, done right. 

const stack = {
  frontend: 'React',
  backend: 'Node + Flask',
  ai: 'PyTorch + LLMs',
}

Cloud infrastructure

AWS, GCP, Terraform, containers at scale. Infrastructure that doesn't page you at 3am.

AWS GCP Docker Terraform Cloudflare CI/CD

// projects

Things I've built

Open source tools and research projects. Each lives on its own subdomain.

mcp-scanner.atharv.io// featured

MCP Security Scanner

Automated vulnerability detection for Model Context Protocol servers. Finds prompt injection vectors, auth bypasses, and data exfiltration risks.

PythonLLMSecurityMCP
redteam.atharv.io

AI Red Team Toolkit

Framework for systematic adversarial testing of language models. Jailbreak detection, guardrail assessment, and automated reporting.

PythonPyTorchRed Team
threats.atharv.io

Threat Model Visualizer

Interactive threat modeling tool for AI/ML systems. Maps attack surfaces, identifies risk vectors, generates STRIDE analysis.

ReactD3.jsSecurity

// experience

Where I've been

Principal Engineer

@ Palo Alto Networks

Security for AI / Prisma AIRS

Leading AI security research. Red teaming language models, building security tooling for the Prisma AIRS platform.

Engineer

@ Protect AI

Acquired by Palo Alto Networks

Built AI security products. Contributed to threat research, vulnerability detection, and model security assessment tools.

Engineer

@ Sydelabs

AI security research and development. Worked on adversarial testing frameworks and security evaluation pipelines.

Engineer

@ MPL (Mobile Premier League)

Full stack engineering at scale. Built and maintained high-traffic consumer-facing applications.

// let's talk

Build something together?

Always interested in AI security research, open source collaboration, and interesting problems.