Hackathon Showcase
SkillVault: The Compounding LLM Wiki
Team consisting of MSc AI and Cloud Computing graduates from DCU and NCI with software engineering backgrounds at Tejas Networks and EY, skilled in C++, Go, Python, AWS, and Azure.
Project Description
Resumes are fundamentally broken for students and early-career builders. People do a lot of practical work—part-time jobs, hackathons, and side projects—but they forget the details and struggle to translate those raw experiences into corporate competencies. Standard AI resume builders just hallucinate keywords from scratch.
SkillVault fixes this by acting as a compounding, stateful knowledge base (an LLM Wiki). Instead of one-off generations, the user drops a messy, informal brain-dump of their latest work shift or project into the app.
Team
Products & Tools
ATS-ready output in seconds.
Anthropic Claude
Autonomy: The agent acts as an autonomous Wiki Maintainer. It independently performs two operations: INGEST (updating the master wiki state with new skills) and QUERY (fetching from that updated state to generate tailored resume artifacts).
Clarity: The user experience requires zero technical skill. A simple drag-and-drop of a raw text log instantly yields a visually striking
Core Engine: Lyzr AI Architect and Lyzr Studio for multi-agent orchestration.
Craft: The workflow relies on advanced state management and data extraction
Database: MongoDB (Lyzr Studio built-in DB) for stateful memory management (storing and retrieving the user's master wiki).
Execution: The pipeline runs flawlessly end-to-end
Frontend: Lyzr Theme Engine (Heritage Dark) and Markdown parsing for a premium
LLM: Anthropic Claude (via Lyzr) to handle complex extraction and formatting.
Lyzr AI Architect
Lyzr Studio
Markdown.
MongoDB
Products & Tools Used
Technical Stack & Frameworks:
Usefulness: It solves a massive
high-contrast dashboard UI.
markdown-formatted dashboard.
moving far beyond a basic text-generation prompt.
permanent record of their actual skills rather than starting from a blank page every time they need a resume bullet.
taking unstructured text and returning highly structured
universal pain point. Users build an ongoing
Additional Links
Summarizing URL...