Projects

Production-grade AI systems I've built — from self-correcting data pipeline generators to NLP-driven cybersecurity tools.

AI-Powered Data Pipeline Generator

Built an end-to-end GenAI web app that converts natural language prompts into production-ready Microsoft Fabric data pipelines using Google Gemini 2.5 Flash.
Designed a self-correcting Agent Mode with automated validation loops and schema-aware code generation — auto-fixing pipeline errors across up to 5 iterations without human intervention.
Implemented real-time SSE streaming so users see pipeline generation live, paired with Azure AD OAuth 2.0 and secure token-based session handling for enterprise-grade authentication.
Reduced data pipeline development time by 80%+, enabling non-technical users to create complex data transformations in minutes.

AI-Driven Vulnerability Detection & Testing System

Engineered an AI-powered cybersecurity dashboard using Flask and the Google Gemini API that automates real-time vulnerability detection, CVE analysis, and SQL injection scanning.
Integrated the Vulners API to fetch 100+ daily CVE entries and leveraged Gemini to perform NLP-based threat analysis — extracting exploitable weak points and attack vectors from raw CVE data.
Built an automated SQL injection scanner using BeautifulSoup and payload-based form testing across 5 injection vectors, with structured JSON/text reporting for each scan.
Developed a responsive security operations dashboard (dark/light theme toggle, real-time API interactions) designed for professional security workflows.