UW LifeBridge

Inspiration

International students often receive important university emails about tuition, immigration, CPT/OPT, deadlines, forms, events, and appointments. The information is useful, but it is easy to miss, misunderstand, or forget the next action.

What it does

UW LifeBridge is an AI assistant that helps international students understand university communications and manage follow-up tasks. It identifies important information, summarizes messages, extracts deadlines, suggests next steps, and surfaces only decisions that require the student's attention.

How we built it

The agent uses the Strands Agents SDK to coordinate specialized tasks such as email understanding, deadline extraction, action planning, and official-link verification. The system is designed to work autonomously while keeping the student in control of important decisions.

Challenges we ran into

The main challenge was distinguishing urgent action items from general announcements and presenting information clearly without overwhelming students. We also focused on making the agent careful with sensitive academic and immigration-related information.

Accomplishments that we're proud of

We built a safe human-in-the-loop workflow that converts a confusing university message into a clear, actionable checklist. The MVP also supports a demo mode without AWS credentials and a live Amazon Bedrock mode when configured.

What we learned

We learned that a useful AI agent should not only answer questions. It should understand context, track tasks over time, and know when to ask for confirmation instead of acting automatically.

What's next for UW LifeBridge

Next, we plan to add secure email integrations, calendar reminders, multilingual support, and verified university source retrieval. These features will remain permission-based and will never take sensitive actions without student approval.

Built With

• amazon-bedrock
• amazon-dynamodb
• aws-lambda
• aws-ses
• python
• strands-agents-sdk
• streamlit