AI-Powered Security Awareness Platform

A production SaaS platform that helps companies train employees against modern social engineering attacks — vishing calls, AI-powered phishing, and deepfake impersonation. The platform runs realistic simulated attacks using AI voice agents and measures how employees respond to each attempt.

Owned the vishing and phishing modules end-to-end — backend and frontend — plus integrating the deepfake service into the broader platform. Also owned the production readiness on these modules: edge cases, error handling, and the operational side.

• Built the multi-tenant campaign engine: admins configure AI agent prompts, select cloned voices via ElevenLabs, provision Twilio phone numbers, and schedule vishing simulations — all backed by Celery for reliable job execution with retries and dead-letter handling.
• Integrated Vapi for AI voice agents — the platform sends full campaign configuration (agent prompt, cloned voice, target phone, caller ID) and receives transcripts, recordings, and evaluation scores via webhook for per-employee outcome reporting.
• Designed and built a Chrome extension in React that records audio directly from any browser tab — enabling customers to clone voices from real meetings on Google Meet, Slack, or any web-based call platform, eliminating the manual download-and-upload workflow.
• Solved the Chrome tab audio capture constraint — built a session-management layer in the extension's background script to track capture bindings across tab switches, with clear visual feedback so users always know which tab is being recorded.
• Built the customizable email builder and SMTP injection layer for phishing campaigns, bypassing spam filters via tenant-side mail configuration.
• Implemented workforce synchronization with Okta, Slack, Microsoft, and Google for employee directory management across tenants.

Platform Overview

Enterprise security awareness platform with AI-driven vishing simulations, phishing campaigns, and deepfake training modules.