RETHICK CB@rethickcb
2
Events attended
2
Submissions made
India
2+ years of experience
About me
Motivated Computer Science and Data Science student passionate about Artificial Intelligence, Machine Learning, Geospatial Technology, and building real-world technology solutions. Interested in solving practical problems through innovation, intelligent systems, and scalable software development. Continuously learning and exploring modern technologies including AI, Data Science, MLOps, and software engineering while working on impactful projects and collaborative development.
🤓 Latest Submissions
PROMETHEUS SHIELD
PROMETHEUS SHIELD is a self-healing AI security platform that protects enterprise AI agents from prompt injections, jailbreaks, data leaks, credential theft, and other AI-native attacks in real time. The platform acts as a security middleware layer between users and AI agents. Using Lobster Trap DPI and Gemini AI, it detects malicious prompts, analyzes vulnerabilities, auto-generates YAML security policies, and permanently blocks repeated attacks through an autonomous defense loop: Attack → Detect → Analyze → Patch → Replay → Block. For demonstration, we built a protected ecommerce AI assistant called “ShopEasy AI Assistant,” where PROMETHEUS SHIELD successfully detects attacks, deploys live patches, and improves security scores in real time through a futuristic cybersecurity dashboard. The platform includes: Autonomous Red-Team Attack Agent Lobster Trap Security Layer Gemini-Powered Vulnerability Analysis Auto-Patch Deployment Engine Real-Time Threat Dashboard Audit Logs and Risk Analytics PROMETHEUS SHIELD is designed to become the missing firewall layer for autonomous AI agents. Tagline: “AI That Defends AI.”
19 May 2026
PR Autopilot
PR Autopilot is an AI-powered pull request intelligence platform built to solve one of the biggest engineering bottlenecks: code review risk analysis. Using IBM Bob, the platform analyzes pull request diffs together with full repository context to detect hidden dependency impacts, transitive breakages, and high-risk modules before a reviewer even opens the PR. The core feature is the Blast Radius Graph — an interactive D3.js visualization that maps directly changed files, transitively affected modules, external dependencies, and safe untouched services in real time. Developers can click any node to view IBM Bob’s reasoning for why a file or service is considered at risk. PR Autopilot integrates with GitHub webhooks, automatically triggering analysis whenever a pull request is opened. The backend, built with FastAPI, fetches repository context and PR diffs, sends them to IBM Bob for repository-aware analysis, then generates dependency graphs, risk scores, changelog drafts, and automated PR summaries. The platform helps engineering teams reduce review time, prevent production outages, and improve confidence in large-scale code changes. By transforming static PR reviews into visual AI-assisted impact analysis, PR Autopilot introduces a new workflow for modern software engineering teams.
17 May 2026
