The provided text is well-structured and clear. There are only a few minor suggestions for improvement to enhance flow and precision.

Here's the fixed version with explanations for the changes:

Inspiration

In my family, I have heard the complaints "hot flashes", "painful legs", "difficulty to sleep" (in Moroccan Darija: sahda, rejlya tay7er9ouni, matayjinich n3ass) several times throughout the years. At first, it was my grandmother who complained about them, then later my aunt, and then my mother. During my PhD, I had the chance to attend a talk by an expert on cervical cancer, who mentioned that most chronic diseases for women are related to their hormones, lifestyle, and beliefs. This pushed me to investigate further into commonly experienced women's diseases and hormonal changes, which is how I understood what the older women in my family were experiencing—menopause. It also helped me understand what could best help them go through it smoothly. – by Sara El-Ateif

Menopause is a natural process, defined as "the ceasing of menstruation" (source: Oxford). Every woman goes through it; some have an easier time, while others suffer various health issues due to different causes primarily related to nutrition, stress, and lifestyle choices, such as hot flashes and cardiovascular diseases.

Every woman has different issues based on her body type and lifestyle, making a single solution challenging. Therefore, a personalized program is needed to help these women overcome their specific challenges. Most importantly, it is crucial to listen to these women and not downplay the symptoms they are experiencing.

What it does

Our solution, MenoGuide+, aims to offer women experiencing difficult menopause personalized nutrition, emotional support, and connections to the community—a solution tailored specifically for them.

How we built it

We used the Google Agent Development Kit to define three specialized agents:

  • Agent 1: Nutrition Expert: Has extensive knowledge of how to alleviate menopause symptoms through targeted healthy food consumption.
  • Agent 2: Life Coach: Acts as a coach, helping users deal with emotional outbursts or stress by leveraging life coaching principles (i.e., asking questions that help users identify what is wrong and guide them toward solutions and effective implementation).
  • Agent 4: Community Connector: Finds support groups, facilitates sharing similar experiences, curates content, connects women with similar symptoms, and helps locate local doctors. This agent helps users find other women going through similar experiences and connects them with support groups and doctors nearby.

Tools

  • ADK: Agent Development Kit, used to develop the three expert agents listed previously.
  • Gemini: The LLM powering the Agents in ADK.
  • Google Cloud Storage (GCS): Used to store research papers and blog posts that serve as a knowledge base, empowering these agents with the latest clinically proven strategies on how to deal with Menopause.
  • Vertex AI Search: Used to test and query the database on the stored data.
  • Flutter: The front-end application that the user interacts with.
  • Firebase: Used to store data for the front-end app.

Challenges we ran into

Finding accurate, clinically approved data is challenging. There are few research papers, expert blog posts, and scientifically relevant and diverse information available specifically on Menopause.

  1. Orchestration between agents.
  2. Local configuration of agents.
  3. Lack of data for menopause-specific communities.

Accomplishments that we're proud of

Building an app that will offer women insights into what they are going through and support them through this difficult period. Most importantly, we want to make them feel heard and seen.

What we learned

We learned how multiple agents can replicate different human experts and simulate a team of experts working together, and how accessible today's technology, like Google's ADK, Gemini, and Vertex AI, makes all of this.

Specifically, we learned how ADK works, how agents are deployed, and how to store data in Google Cloud Storage. Additionally, for app development, we learned how to link the AI to the Flutter application.

What's next for MenoGuide+

We plan to test the current solution by making it publicly available and improving the agents' precision by providing more clinically relevant data and learning in real-time from women experiencing similar symptoms.

We hope to make this available to our community, so we will also work on providing a Darija version of the app.

We also plan to develop a Facebook-like community section where women can explore and connect with others who have similar symptoms.

Built With

Share this project:

Updates