Aditya Rai, Mumbai
Ask first.
Then build.
I'm an engineer who has to meet the people a thing is for before I can build it. Below is what they told me, and what I made.
Builds things for fun
Weekend rider
One more anime episode
What people said, and what it became
Every feature here started as a sentence someone said to me. Pick a project and watch the two line up.
MediSync
“Is the doctor even in today?”
Live doctor check-in
RFID, camera and Wi-Fi confirm who's in
“Serious cases wait like everyone else”
Severity-first queue
AI score moves urgent cases up
“Nobody knows which beds are free”
Bed tracker
Free beds visible to staff and patients
Six times I asked first
Each one in the same order: who I asked, what I heard, what I built, and what happened next.
2023
MediSync
Public hospitals that see the sickest patients first
- Who I asked
- Doctors and patients at a government hospital
- What I heard
- Patients often can't tell whether a doctor is actually in, and the most urgent cases wait in the same line as everyone else.
- What I built
- Four connected portals for government, hospitals, doctors and patients. Doctor attendance is confirmed by RFID, camera face recognition and Wi-Fi, and an AI severity score moves the sickest patients up the queue.
- What happened
- National finalist at Smart India Hackathon 2023, top 6 of 500 teams.
Read the MediSync storyCode on GitHubMade with Python, AI severity scoring, RFID, Computer vision, Web portals
Try it: the OPD queue
Sample data
Switch the queue from arrival order to severity and watch who gets seen first.
Dr. Rao: in since 09:12, confirmed by RFID, camera and Wi-Fi
- 1Routine BP checkPatient 1, arrived 08:052/10
- 2Prescription refillPatient 2, arrived 08:121/10
- 3Fever for 3 daysPatient 3, arrived 08:205/10
- 4Chest pain, breathlessPatient 4, arrived 08:319/10
- 5Sprained anklePatient 5, arrived 08:403/10
Seen first: Routine BP check, because they arrived first.
2023
withU
One tap, and the people nearby know
- Who I asked
- 14 women about the moments they felt unsafe
- What I heard
- Calling for help takes too many steps, and the people who are physically closest never find out.
- What I built
- A React Native app with one-tap SOS that alerts emergency contacts and app users within a few kilometres, shares live location, and shows the nearest police stations and hospitals.
- What happened
- 2nd place among 150+ participants at TechSparks (FCRIT).
Try it: one-tap SOS
Sample data
Press and hold the button (or press Enter). Everyone inside the circle hears about it.
- 3 emergency contacts saved
- 5 app users within 3 km
- Nearest police station 900 m, hospital 1.4 km
2025
Co-Life × SaafWater
Making an invisible water crisis visible
- Who I asked
- SaafWater's founders, an IAS officer who buys water infrastructure, and a water treatment site in Goa we visited
- What I heard
- India's water crisis is hidden: many villages test water monthly or not at all, lab results take up to a week, and the officials who fund water systems hadn't heard of the startup that monitors it live.
- What I built
- Impact metrics that turn real-time water data into health, cost and farming numbers, and a brand-awareness plan that uses those numbers as talking points, repositioning SaafWater from a tech startup to an impact business.
- What happened
- Delivered the impact metrics and an outreach plan to SaafWater, a startup featured on the Prime Minister's Mann Ki Baat, and completed the Co-Life impact entrepreneurship programme.
Read the Co-Life × SaafWater storyMade with Design thinking, Stakeholder interviews, Site visit, PESTEL, VRIO, Impact metrics
Try it: poison the tank
Illustrative network
Pick who's watching the water, then contaminate the tank and see how far it gets before anyone knows.
00:00Homes reached: 0 of 18
2025
Saint-Gobain Solar Gard
46 locations on foot to size a ₹6 crore market
- Who I asked
- 40+ people in the Delhi NCR sun-film trade, from detailers to dealers, across 46+ locations
- What I heard
- Car sun films are a fragmented, semi-organised market where buyers struggle to tell brands apart at the shop counter.
- What I built
- Tableau dashboards on competitor import data (top products, pricing, demand), product spec recommendations, a distributor shortlist and a go-to-market plan for entering automotive films.
- What happened
- Sized a ₹6 crore opportunity, generated ₹3 lakh+ in qualified leads on the ground, and the company adopted the plan and launched in 2026.
Read the Saint-Gobain Solar Gard storyMade with Field research, Stakeholder interviews, Tableau, Import data, Go-to-market
Try it: walk Delhi NCR
Real totals, sample split
Tap each area to survey it. The market picture builds up as you go.
- Locations
- 0
- Interviews
- 0
- Opportunity
- ?
5 areas left to survey.
2025
Bombay Shaving Company
38 stylists taught me what a pro tool needs
- Who I asked
- 38 stylists across 13 cities, plus 120+ surveys and 3,200+ product reviews
- What I heard
- Salon professionals need tools that last through a full working day and get serviced, not consumer trimmers sold once.
- What I built
- A go-to-market plan for the category team: product design inputs, subscription and rental models, a distributor plan and an influencer channel.
- What happened
- We developed a full go-to-market plan for the professional range, from product design to how salons would buy and service it, and presented it to the category team.
Read the Bombay Shaving Company storyMade with Interviews, Surveys, Review mining, Go-to-market
Try it: from research to plan
Sample data
Open each source. The plan fills in as the evidence comes in.
One-on-one conversations with salon professionals across 13 cities.
The go-to-market plan
- Product design for professionals
- Subscription and rental models
- Distributor plan
- Influencer channel
2025
V-Guard Big Idea 2025
A product idea, priced down to the last component
- Who I asked
- Climate data for every Indian state, household counts, and V-Guard's own annual report
- What I heard
- Indian kitchens get hot and humid, and the hottest, most humid states also hold the most households.
- What I built
- A business case for a compact kitchen cooling and air-cleaning unit: a climate index to choose target states, component-level costing, a ₹6,000 price and break-even at about 3,450 units.
- What happened
- Top 21 of 3,000+ registrations in V-Guard's national innovation challenge.
Read the V-Guard Big Idea 2025 storyMade with Market sizing, Unit economics, Break-even analysis, Financial forecasting
Try it: the break-even
From our financial model
Change the price or the cost scenario. See how many units it takes to pay back the launch.
- Parts cost per unit
- ₹2,678
- Earned per unit
- ₹3,322
Units to break even3,450Fixed cost: ₹1,14,60,600, about 2% of V-Guard's yearly fixed costs, from our model.
The days it worked out
Every dot is someone I competed against. The yellow ones are where I finished.
National hackathon2023Top 6
of 500+ teams
Smart India Hackathon
National finalist with MediSync
1 dot = 2 teams
- 36-hour national hackathon
2nd
of 600+ participants
HackOverflow 1.0
1 dot = 2 participants
- Hackathon2023
1st
of 100+ participants
Tech Hack, Amity University
- National innovation challenge2025
Top 21
of 3,000+ registrations
V-Guard Big Idea
1 dot = 10 registrations
- Innovation case competition
1st
of 240+ participants
NetApp Nexus
- National marketing challenge2024
1st
of 250+ participants
PRisma'24, University of Delhi
- 24-hour national hackathon
2nd
of 300+ participants
TechNight
- Project competition2023
2nd
of 150+ participants
TechSparks, FCRIT
With withU
- Technical Lead, Google Developer Student Club: led a 13-person team that ran events for 160+ and 130+ people
Smaller things I've made
App2026
SalesFlow
Finds leads automatically and runs personalised email outreach. FastAPI, SQLite, Gmail API, Playwright.
Research2024
Soul Travelling
Surveyed 70+ travellers and interviewed 10 people across 3 groups; found the 5 factors that drive trail choice.
Web2022
Kaash Light Engineers
Built responsive pages for a Luxembourg lighting-design firm and cut page load time by 34%. The site is live.
App2023
TogetherBridges
Web pages, volunteer sign-ups and Stripe donations for NGOs.
App2023
GradGuide
A money app for students: expenses, gigs and lending in one place.
Web2022
The first code I ever wrote
A university website, built from scratch while I was learning HTML and CSS. Where it all started.
How I work
Go and see
Before writing code, I talk to the people who'll use it. Fourteen conversations for withU. Thirty-eight stylists for a trimmer range.
Write down what hurts
The real problem is usually smaller and more specific than the brief. I highlight it, literally.
Build the smallest useful thing
A working prototype starts better conversations than a deck does.
Go back and ask
Then fix what's wrong and ask again. Most of the good ideas arrive on the second visit.
When nobody's asking me anything
I ride, I travel, I swim, I watch more anime than I admit, and I play one more round of whatever game it is than I planned to.
Weekend rides up into the hills with friends. 
Chasing fog through the tea estates of Munnar. 
Yes, that's an anime tapestry. No, I'm not taking it down. 
Feet in the Ganga at Rishikesh, after the rafting. 
Nikhil, Deep, Aman Agasthya and me at GIM. Two years, one sign. 
Aman and me, losing a fight with the waves. 
Last day of engineering, with the people who got me through it. 
Usually the last one out of the water.
Right now I'm building small internal tools for a life insurer's sales teams. They're under NDA, so there are no screenshots here, but I'm happy to talk about how I approached them.