What the rooms shipped.

Astha Singh

“Gemini was fun”

Jocelyn Gu

LYNX lets you add your whole wardrobe into an AI wardrobe so that you can mix and match your daily outfits, get new clothing suggestions, and also try on new pieces of clothing from online shopping. It's your online wardrobe but with a lot of customization

“What really surprise me was how powerful gemini was, especially in combination with the live video”

Hakeem Fuad

It's the building blocks for the future of product recommendations. Users complete a quick body-profile questionnaire to generate a customizable 3D mannequin, view estimated measurements, explore clothing options, and receive size recommendations from a sample garment catalog. It is designed to make online shopping feel more personal by showing how body shape and fit preferences affect garment selection.

“Great team. Great experience.”

Jie Peng

SecondLook makes fashion resale as simple as showing an item. Instead of starting with a listing form, it uses multimodal AI to see, listen, understand condition, generate the listing, research the market, and recommend a price. By removing the friction of resale, SecondLook helps fashion move from one person’s past into another person’s style. https://drive.google.com/file/d/1M-gB_qWJuZUL2FXceOGHevAX9Xb6tVnU/view?usp=drive_link

“Time is a little tight!! Lol, but it is FUN!!”

Anjana Modugu

“i was surprised with how fast the time went by”

Jocelyn Gu

an AI stylist that can see your closet, and see you. We built an AI co-shopper that actually knows your real closet, because it scanned it. Hold any item up to a live camera: shirt, shorts, shoes, doesn't matter how many at once, and Gemini catalogs it into your digital closet in seconds, no manual photo uploads. Tap any item and the stylist immediately gives you feedback.

“how powerful gemini was”

prakhar tripathi

Everyone already does this. You're standing in front of a mirror before something that matters, you take a photo, you send it to three friends, and you wait. Then you get back "yeah nice" from one and nothing from the other two. Committee Couture puts the picture in the middle of that conversation. You get on a call with your friends, one of you sits in frame, and the others just talk. Say "green french coat with black buttons" and it's on them. Say "beige beret" and that's on them too. Every su

“The vonage API was giving us some issue, but thanks ot the team present at the hackthon we were able ot ship in time”

Putta Sathvik

Deadstock Live Lab is a material-grounded AI co-design studio for circular fashion ateliers and independent designers. Instead of hallucinating garments from thin air, designers scan physical fabric leftovers and rolls using live camera vision, and Gemini instantly estimates yardage, weave, and drape to synthesize a manufacturable capsule collection. The cool part: every garment zone is mathematically bound to observed deadstock inventory, and if a roll is removed mid-session, the causal constra

“What surprised us was Gemini Flash's ability to reason about physical textile physics directly from camera frames. It correctly identified weave structures, estimated fabric weight/drape, and understood that a rigid denim remnant can anchor an overcoat shell while a lightweight offcut should be reserved for collar bands or sleeve linings. Furthermore, when paired with structured schema and token c”

harshita patil

“nothing really”

Maria Mejia

“The capabilities of front end”

Shagun Sharma

“abc”

Aaditya Prabu

Pocket Stylist is a personal stylist on your phone. It looks at your body and learns your personality, then shows you wearing outfits made for you. You take four photos (front, back, left, right) and one of your hand, then add your height, weight and a short description of yourself. Pocket Stylist then: 1. Measures your body on your device. Computer vision (MediaPipe pose estimation and body segmentation) estimates your shoulder width, chest, waist and hips, arm length, inseam, and body shape.

“I expected to give some time but it was a very short, we werent able to develop the frontend”