Recently, I’ve been working with a friend of mine on building a lifestyle brand that currently focuses on clothing, skincare, and related food habits.

Started working on the skincare POC,
where we first understand the user’s skin. For the POC, we fetch a user’s image and share it with an LLM to get a skin review. This review covers skin color, skin type, and other basic skin concerns (like acne). This review is then sent to another LLM call, which returns a list of recommended products for each skin concern.

In the future, the plan is to find the right questions to ask the user to get more details. These additional details, along with the image, will be sent to an LLM. And obviously, instead of just asking the user to upload an image, a better camera option will be added.

Coming to product recommendations — we don’t want to just throw some products at the user. We want to give them a proper routine and identify the products they’re already using (mostly by reading the image of their products), then suggest new ones if they’re missing out. With this data, we can curate a personalized routine and give it to the user with edit options.

Once the routine is fixed, we can provide a tracker for the user to follow up on their routine with a reward system.

Along with this, lifestyle changes like food intake will also be recommended.

This is just the start. We want to make the skincare module production-ready first, then move on to other modules like clothing and food habits.

Happy to have my friend, Sadhana Das, helping me with analyzing the product, giving ideas, and reviewing my thoughts.