Poshan
Poshan ("the kitchen that remembers") is an academic concept set in Silchar. A shared "household model" of members, health goals, pantry and local supply quietly helps the person who cooks: it puts meal options and recipe steps on the kitchen wall, sends the grocery list on WhatsApp, and phones the grandparent who'll never open an app. It sits alongside a calorie-tracking companion app, which the project labels its prior work.
UI/UX
Product Design
Role
[TBD]
Timeline
[TBD]
team
Guided by Prof. Anirudha Joshi
platform
Mobile
What this is for
When a doctor says “less salt” or “watch your sugar,” the advice goes to one person. But in many Indian homes one person cooks for everyone, with whatever is in the kitchen that day. Most nutrition apps ask each eater to log every meal. Many people stop.
Poshan designs for the household instead. One shared picture of the family (who needs what, what’s in the pantry, what the local market has, what got cooked and what came back uneaten) feeds help into the kitchen itself, into WhatsApp, and into a simple phone call. The deck opens with one line: “Most nutrition plans fail in the kitchen.”
Process (how the work was paced)
This process is reconstructed from the slide labels and screens. [YOU: correct the order if needed]. It started with prior work: a full calorie-counter app (scan a meal, see nutrients, streaks, family groups), already localised with tabs like At Home · Office Canteen · Zomato. Next came observation of nutritionist WhatsApp groups and interviews, which pointed away from more logging and toward context. A first framework followed: map the household’s food world, co-create a few options rather than one diet, test one small change at a time, and “if the barrier is structural, stop counselling it.” Then the system: “Six surfaces. One household model.” (kitchen cameras, a kitchen computer, voice, WhatsApp, AI calls, the app), arranged as a ladder from camera-plus-projection down to a keypad phone. A run of time-stamped storyboards, from 6:15 AM to 7:40 PM, tests the idea across a day. An Ethics slide and a planned pilot close the deck. Stage 1 of P1, guided by Prof. Anirudha Joshi. [YOU: timeline, role]
One household, many doors
The user is the household, not a person with a phone. So Poshan keeps a single model of the family and lets each member reach it in their own way. The kitchen camera and projector are the most accurate layer. The companion app is for people who’ll self-log. WhatsApp needs no install. An AI phone call reaches the elder with no smartphone (“Can be a key pad phone”).
Suggestions carry everyone’s needs at once. A morning strip of five dishes is tagged with why each one fits: “safe” for the child, “low salt” for the grandmother, “low GI” for the father, “uses kal ka bhat” (yesterday’s rice), “12 min.” She points and says one word. “Five options. One word. No scrolling.”
Information arrives where the decision is made
Help appears at the moment it’s useful, without asking for input. Recipe steps are projected on the wall while her hands are in dough, including swaps (“curry patta nei — skip, taste holds”). A nudge reminds her the green peas are on the fridge’s second shelf. Ask by voice for the child’s lunch and it offers only dishes he has actually finished before. In the evening the kitchen notices what’s running low and what keeps coming back uneaten, then sends a WhatsApp list or fills a delivery cart. Captions from the file: “It stopped buying what he never ate.” At breakfast everyone’s progress shows on the table: “Nobody was told what to eat. Everybody could see.”
The biggest open question is how the kitchen knows what it holds. Cameras imply it, but reading a closed fridge, correcting mistakes, homes without a camera, and consent aren’t designed yet.
The challenge
The person who’s ill usually isn’t the person who cooks. In the deck’s Silchar story, a 72-year-old has high blood pressure, a 45-year-old is watching his blood sugar, a 12-year-old eats “taste only,” and the woman cooking for all of them has “no plan of her own.” On top of that, the evidence in the file says logging fails: from about a month watching three people’s nutritionist WhatsApp groups, the deck notes that “detailed food logging increases burden and is commonly abandoned” and that tracking can bring “guilt, restriction and anxiety.” Interviews add the rest: some foods are “impossible, not preference,” and availability depends on money, floods and ration shops. [YOU: number and type of interviews]
Status (honest) and next
What’s real today: Poshan is a Stage 1 concept presentation in Figma, with a 21-slide research and storyboard deck plus a high-fidelity companion app of 74 screens in both Light and Dark, 37 more exploration screens (an earlier “My P1” onboarding set, 13 dashboard variants, three date-selector styles), and a documented token system (primitive → alias tokens with Light/Dark values, Inter + Manrope type, accent blue #1570EF on near-black #0A0A0A). The app still counts calories, which sits uneasily with the deck’s own finding about guilt and labelling. Nothing has been tested with families. The pilot (app + WhatsApp bot + calling agent, a nutritionist consult, and GSQC / ICMR-NIN “plate of the day” measures for an intervention study) is planned, not run. [YOU: built / prototyped? jury feedback you can quote?]
Next: run the pilot, design how pantry inventory is captured and corrected (and what happens without a camera), settle camera consent and data ownership, and spec the projected UI and WhatsApp/call scripts.
Publishing note: before going public, confirm whether the Silchar family is a composite, credit how the storyboard art was made, and confirm which app screens and tokens are your own vs. adapted from a kit. [YOU: confirm role, timeline, and team]
