Founder at RokaMaybe · Building a private tracker for the Indian arranged marriage search

Komal Singhania • Peerlist

<h3>RokaMaybe is a private arranged marriage tracker designed to streamline your rishta search. It leverages AI to read biodatas, automatically computes detailed 36-point Ashtakoot kundli matches based on Vedic astronomy, and provides a centralized dashboard to log calls, meetings, and red flags. Key features include an AI biodata reader that populates profiles from uploaded documents or photos, a prospect tracker with a five-stage journey (new, call, meet, met, decision), a call journal with a timeline, a red-flag and green-flag note system that is searchable, and a side-by-side comparison tool for shortlisted prospects. It also offers a 'Mummy Mode' for sharing a sanitized view of the search with family and allows for partner preference settings and match scoring. The platform emphasizes privacy, ensuring data is secure, encrypted, and only accessible by the user, with options to export data or delete the account at any time. RokaMaybe is a web application accessible via modern browsers on any device, offering a free early access period with all features unlocked.</h3><ul><li>AI Biodata Reader: Upload biodatas (PDF, Word, photo) and let AI fill in the details automatically.</li><li>36-Point Kundli Matching: Accurate calculations based on Vedic astronomy, flagging key doshas.</li><li>Prospect Tracker: Manage your rishta search through a clear five-stage journey.</li><li>Call and Meeting Journal: Log interactions and notes with a running timeline per prospect.</li><li>Red/Green Flag System: Record and easily search important observations about prospects.</li><li>Side-by-Side Comparison: Evaluate shortlisted prospects across multiple dimensions.</li><li>Mummy Mode: Share a curated view of your search with family.</li><li>Partner Preferences: Set and score matches against your desired criteria.</li><li>Privacy Focused: End-to-end private, no ads, no data selling, exportable data, and one-click deletion.</li></ul>

RokaMaybe — Your Arranged Marriage Tracker

Kairo is a full-stack SaaS application designed to solve one real productivity challenge professionals face: tasks hidden inside emails that get forgotten, skipped, or missed.

Using Google’s Gemini 2.5 Flash, Kairo intelligently reads Gmail messages, identifies action items, understands urgency cues, extracts deadlines from natural language (“by Friday”, “EOD”, “next Monday”), and instantly converts them into structured tasks inside a drag-and-drop dashboard.

The system automatically polls Gmail every few minutes using background services, prevents duplicate task creation, maintains a user-specific task ecosystem, and gives professionals a clean, centralized view of everything they need to deliver.

From AI-powered task extraction to a real-time Kanban experience, Kairo brings together AI + productivity + automation into a single, powerful workflow tool.

My Role 

- Designed and developed the complete end-to-end architecture for Kairo.
- Built the AI pipeline using Gemini 2.5 Flash for task extraction, priority detection, and deadline parsing.
- Integrated Gmail API with OAuth 2.0 to securely fetch and process user emails.
- Developed a background polling service to continuously monitor inbox activity.
- Implemented a user-friendly Kanban dashboard with drag-and-drop task management.
- Engineered the backend using Flask with modular services (Gmail service, Gemini service, polling service).
- Created a clean, responsive frontend using React, Tailwind, and modern UI animations.
- Ensured security compliance through JWT authentication, CORS implementation, and encrypted token storage.
- Documented the system with full technical specs, architecture diagrams, API docs, and setup instructions.

Kairo - AI That Converts Email Conversations 

Built at the Google Build and Blog Marathon, hosted at the Google office in Hyderabad in November 2025. Kairo is an AI-powered workflow automation tool that converts Gmail emails into structured tasks automatically, using Google Gemini 2.5 Flash for natural language processing. The app monitors a user's inbox in real time, extracts action items and deadlines from email conversations, assigns priorities, and organises everything into a Kanban-style dashboard. It was built on Google Cloud Platform using React, Flask, MongoDB, Gmail API, and Google OAuth 2.0. My role: product strategy, user problem framing, content, and go-to-market narrative. I worked in a two-person team where I led the product thinking and my teammate handled the development. We also published a technical blog on Medium documenting the build. GitHub: https://github.com/bannu773/Kairo_Complete_App Blog: https://medium.com/@komal.singhania1222/kairo-the-ai-that-remembers-everything-so-you-dont-have-to-ee7ebcd88bda Demo Video: https://www.youtube.com/watch?v=Evc3rX9Ag-s