About Devansh
Devansh, 9th Grader
Cohort Experience
What surprised me most was how much emphasis the program placed on reasoning, decision-making, and handling ambiguity rather than just technical implementation. The open-ended nature of the problems pushed me to think more critically about trade-offs, assumptions, and real-world constraints, which made the learning experience deeper and more impactful than I expected.
About MailMate
Project Description
MailMate is an intelligent, AI-driven email management dashboard designed to solve the problem of "inbox fatigue." It connects directly to Gmail via OAuth and uses Google's latest generative models (Gemini and Gemma) to transform the traditional inbox into a proactive assistant. The project features automated spam detection, sentiment analysis, AI-powered summarization, and an accessibility-focused "Read Aloud" feature that cleans up complex email formatting for natural speech.
MailMate - Project Showcase
Devansh identified a universal productivity killer: the chaotic state of modern email inboxes. His inspiration was rooted in his own struggles with time management and the fear of missing critical communication regarding college applications.
Personally I'm not that good with time management... there are too many emails in my inbox I don't know like if if I'm getting this college emails or if like it's actually important or not.
MailMate - Project Showcase
Devansh built MailMate, an AI-powered email assistant designed not just to read emails, but to understand them. Unlike standard email clients, MailMate utilizes Sentiment Analysis and Urgency Detection to tell the user how to feel about an email and when to act on it.
Key Features & Technical Implementation:
• Tone & Urgency Detection: The AI flags emails as "Urgent," "Positive," "Negative," or "Neutral."
• Smart Categorisation: It automatically sorts emails into labels like Work, Finance, and Promotions without manual setup.
• Universal Access: Integrated Text-to-Speech allows visually impaired users to listen to their emails.
Tech Stack
He utilised Python (Pandas) for the backend and Gemini 2.0 / Gemma APIs for the AI logic. For the frontend, he moved away from low-code tools to build a custom interface using React, HTML, CSS, and JavaScript.
Hear it from Devansh
Devansh’s journey required a significant leap in technical capability. His previous experience was limited to Streamlit (which handles UI automatically), whereas this project required building a professional UI from scratch.
The first cohort was like the first step but the advanced cohort was like five more steps advanced... I also got to learn how does AI work internally... what are the networks, neural networks... it just went more deeper.
Overcoming Roadblocks: The development process was not without frustration. Devansh struggled significantly with API rate limits, requiring him to configure "tens of hundreds of API" keys just to ensure the system ran smoothly during testing. However, the result was a system that the evaluator noted looked "very professional" and functioned in real-time.
Parent Experience
I would definitely just currently out to look at it... maybe I can share it with my community in terms of how to use it and then take feedback from them as well. We have a lot of users in the corporate world.
When Devansh demonstrated Mailmate live, showing the AI analyzing "Work" emails and detecting urgency, his father was stunned by the utility and the complexity of what his son had built.
Project Timeline
