Kairos

• 

  Kairos

• 

  Dashboard

• 

  Opportunity Scanner

Inspiration

As a student who is passionate about technology, I enjoy participating in hackathons, events, internships, and any opportunity that can boost my career. The challenge? All those opportunities were scattered across Notion pages, Google Docs, spreadsheets, and browser bookmarks. Keeping track of deadlines, milestones, and progress for each became a logistical mess. I needed one central hub for my entire competitive career pipeline, which is why I built Kairos.

What it does

Kairos is a comprehensive dashboard that helps you track every career-related opportunity including hackathons, jobs, grants, events, internships all in one place. It gathers deadlines, milestones, checklists, time logs, and even AI-powered preparation guides so nothing gets missed. Rather than juggling multiple note apps, you log an opportunity once, and Kairos keeps everything visible: when the deadline is, what your next milestone is, how many hours you've invested, and what you still need to prepare.

How we built it

Kairos was built entirely on the MeDo platform. We began by sketching the core data model (opportunities, context stacks, milestones, time logs, and AI guides), then developed the React, Tailwind, and shadcn/ui frontend, along with a Supabase-powered backend. The scanner was the most complex part; we created a three-tier fallback system (direct fetch, web reader, web search) that effectively handles sites that block scraping, allowing users to input a URL and receive structured data in return.

Challenges we ran into

The MeDo platform is powerful but operates with a structured workflow. I mention a problem and then fix it, which doesn’t always feel like a natural conversation. Sometimes I had to go to another model to ask follow-up questions or do research before applying changes. Building the opportunity scanner was also challenging: many modern sites like Devpost and LinkedIn are heavy on JavaScript and block direct scraping. We addressed this with multiple fallback strategies, but getting the error experience right (clearly explaining to users why a URL failed and how to paste the content instead) required several revisions.

Accomplishments that we're proud of

• Three-tier scanner fallback. The ability for the scanner to fallback through direct fetch, web reader, and web search when a site blocks scraping is something we're very proud of. It makes the tool reliable in the real world where not all URLs behave well.
• Full AI briefing system. Clicking "Prepare" generates a contextual briefing that includes notes, checklists, milestones, and time logs that provides a complete snapshot before an interview or submission.
• Unified timeline and calendar views. Users can switch between a grouped timeline and a color-coded calendar, giving them two effective ways to visualize their workload.

What we learned

Building Kairos clarified the difference between a simple wrapper and a contextual AI system. Context recall isn’t just about summarizing notes; it involves integrating deadlines, tracked hours, and unfinished checklist items into a predictive briefing. I also learned how to use MeDo's multi-turn generation to quickly iterate on complex UI changes and system performance, transforming instructions into visual timelines, calendar grids, and system updates in just a few days.

What's next for Kairos

• CV builder context layer. We plan to use the profile data and opportunity history in Kairos to create an intelligent CV builder that highlights relevant projects, skills, and outcomes automatically.
• Email reminders. We will send proactive reminders for deadlines and milestones straight to your inbox, so you never miss a submission window.
• Calendar integrations. We aim to offer native syncing with Google Calendar and Outlook so that deadlines and milestones show up where you already organize your time.
• Advanced AI Analytics. We want to recognize outcome patterns to analyze your data and find out which application strategies produce the best conversion rates based on effort versus results.
• Team Collaboration. We’re designing multi-user workspaces specifically for hackathon teams to share context, assign tasks, and manage joint projects.

Built With

• baidu
• ernie
• medo
• node.js
• react
• tailwind
• typescript