AI Health Task Agent

Autonomous AI agent that tracks health, manages tasks, and acts proactively using Gemini and Google Cloud Agent Development Kit

Comment

Inspiration🏥 AI Health Task Agent — Turning Symptoms into Clear Tasks in 2 Seconds

Problem: People Google symptoms and panic with no clear next steps. Solution: Enter symptoms + age group + duration, get instant triage and a 3-day action plan.

How it works — Agentic Workflow powered by Groq openai/gpt-oss-20b:

  1. Analyzer Agent: Parses raw symptom input
  2. Triage Agent: Assigns urgency (self-care / see doctor / urgent)
  3. Planner Agent: Generates home care + task checklist

Demo: Live App: https://9proe.streamlit.app Video: https://youtube.com/shorts/OMxxRE2S9Zo?si=jngnGJ9rT0IJFhtU GitHub: https://github.com/wycliffe36/AI-Health-Task-Agent

Tech Stack: Groq for sub-2s inference, Streamlit for UI, all in one Python file.

Built in Nairobi for All Things Agentic Hackathon.

⚠️ Disclaimer: For information only, not a medical diagnosis. Always consult a healthcare professional.

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for AI Health Task Agent

Inspiration

Millions struggle to remember medication, track symptoms, and manage daily tasks. Existing chatbots only answer questions — they don't act. We wanted to build a true agent that takes initiative and removes friction from daily health management, especially in underserved regions like Kenya.

What it does

AI Health Task Agent is an autonomous agent built with Gemini and Google Agent Development Kit (ADK):

  • Tracks medication schedules and health routines proactively
  • Monitors symptoms and logs health data automatically
  • Manages daily tasks, calendar, and smart reminders
  • Takes real action using tools - not just chat
  • Works asynchronously in the background while user lives their life
  • Provides personalized insights from health trends

How we built it

  • Gemini 2.0 Flash for reasoning and multimodal understanding
  • Google Agent Development Kit (ADK) for agent orchestration
  • Google Cloud Run for deployment and scaling
  • Python, Vertex AI, and custom tools for calendar, notifications, and data logging
  • Strands-style tool calling for autonomous decision making

Challenges we ran into

  • Making the agent truly autonomous vs just a chatbot
  • Handling sensitive health data securely and privately
  • Tool orchestration for multi-step workflows
  • Ensuring reliability for critical health reminders

Accomplishments that we're proud of

  • Built a production-ready agent that plans, reasons, and acts independently
  • Background task execution without constant user prompting
  • Clean integration with Google Cloud ecosystem

What we learned

  • How to design agentic workflows with ADK, memory, and tool use
  • Building agents that are helpful, trustworthy, and proactive
  • Balancing autonomy with user control in health contexts

What's next

  • WhatsApp integration for Nairobi users
  • Voice interface in Swahili and English
  • Integration with hospital systems and wearable devices
  • Predictive health alerts using long-term data

Built With

Share this project:

Updates