
AutoPilot CI is an autonomous, multi-agent CI/CD intelligence platform that transforms every code push into a fully reviewed, tested, and hardened release candidate — without human intervention. At its core, the system orchestrates a swarm of specialized AI agents running in parallel: a Code Analyzer that detects anti-patterns and maintainability issues, a Security Scanner that catches SQL injection, XSS, insecure dependencies, and CVEs, a Performance Profiler that identifies N+1 queries and algorithmic bottlenecks, a Test Generator that writes unit tests for untested functions, a Docker Scanner that validates container configurations, and a Supervisor agent that aggregates all findings and makes the final pipeline decision — auto-fix, escalate, or deploy. When the supervisor decides to act, an AutoFix agent applies patches directly to the codebase, commits them to a new branch, and opens a pull request with a full markdown summary of every change made. A Deployment agent then handles rolling or blue-green release strategy based on severity thresholds. The platform supports Python, C#/.NET, and TypeScript/Angular codebases out of the box. It integrates with dotnet build to catch C# compiler errors statically, respects .gitignore via git ls-files, and offers both diff-only and full-repository scan modes. Every pipeline run streams agent decisions live to a custom real-time dashboard over Server-Sent Events, giving engineers full visibility into what each agent found and decided. At completion, a professionally formatted PDF audit report is generated — covering findings by severity, fix summaries, deployment logs, and a pipeline timeline — ready to share with stakeholders. Built on the Claude Agent SDK with a FastAPI backend, AutoPilot CI demonstrates how orchestrated LLM agents can replace entire categories of manual code review, security auditing, and DevOps toil — making high-quality, secure software delivery accessible at the speed of a git push.
10 May 2026
.png&w=828&q=75)
Procurement AI Assistant is an intelligent tool designed to simplify and accelerate procurement data analysis. Users can upload CSV or XLSX files containing procurement records and interact with the system using natural language queries. Behind the scenes, uploaded files are converted into structured SQLite tables, while a LangChain-powered SQL agent processes user queries with the help of Gemini LLM to generate accurate responses. The frontend, built in Streamlit, provides a simple and interactive interface, while the backend API uses FastAPI to handle requests efficiently. This project enables instant insights, reduces manual effort, and makes procurement data analysis accessible to everyone
23 Nov 2025

AI Study & Research Copilot is a personal AI assistant designed to revolutionize the way students, researchers, and self-learners interact with knowledge. Users can upload PDFs, paste YouTube lecture links, or provide web pages and notes. The system ingests this content, splits it into semantic chunks, and stores it in a vector database (Qdrant) for fast retrieval. Leveraging the Groq LLM, users can chat naturally with their materials, ask questions, generate summaries, create flashcards, and even design quizzes. The AI remembers past interactions, helping build a personalized learning memory. With a clean web UI and modular architecture, it’s hackathon-ready, highly interactive, and scalable, merging RAG, embeddings, memory, and generative AI to transform studying into an intelligent, engaging experience.
19 Nov 2025