The dream of ConnectCatalyst started with a simple yet frustrating reality—sending out hundreds of applications, facing rejection after rejection, and getting nowhere. Many qualified job seekers are stuck in the same cycle, where a simple qualification isn’t enough because networking remains a mystery.
ConnectCatalyst is the breakthrough that levels the playing field, transforming networking from a guessing game into a strategic advantage. Instead of cold outreach and missed opportunities, our Agentic GraphRAG solution pinpoints the right connections—helping job seekers reach the people who can open doors. By exploiting graph intelligence and context-aware introductions, ConnectCatalyst turns weeks of blind applications into warm, high-impact conversations in just days.
Traditional job searching is broken—applying blindly and sending cold messages rarely works. Networking is the key, but most people don’t know who to reach out to or how to get a response. That’s where GraphRAG comes in. By combining graph intelligence with retrieval-augmented generation (RAG), we don’t just match job seekers with companies—we trace optimal connection paths through their network. Who knows someone at your dream company? Who is the best person to introduce you? GraphRAG finds these answers instantly, ensuring every connection is strategic, warm, and high-impact—not just another ignored LinkedIn request.
To power GraphRAG, we needed a scalable, multi-model database capable of handling complex relationships, deep graph queries, and fast retrieval. ArangoDB is the perfect fit—its native graph storage allows us to efficiently map professional connections, uncover hidden networking opportunities, and execute powerful pathfinding algorithms. With AQL and NetworkX integration, we seamlessly combine graph traversal, analytics, and vector search, ensuring every query is optimized for speed, accuracy, and impact.
datasets_generation.ipynb- Generates synthetic LinkedIn network datasets, including LinkedIn profiles, work experiences, education history, job postings, and their relationships.Connect_Catalyst_Final.ipynb- The final submission notebook that integrates the datasets, builds the professional network graph, executes queries, and demonstrates the full capabilities of ConnectCatalyst.
- Some queries were rephrased to improve the demo flow and readability.
- The rephrased queries still work as intended, ensuring that the solution remains functional.
- If the guardrail is too strict, it can be removed during testing for better flexibility.