The Science Behind Terazo’s Top-Five Finish at Summer of Galaxy’s Hackathon

The Science Behind Terazo’s Top-Five Finish at Summer of Galaxy’s Hackathon

Earlier this month, Terazo joined the largest challenge-driven AI hackathon at DeveloperWeek 2024. The event hosted both in-person and virtual participants, with over 600 participants from the international developer community participating in the renowned event.

This year’s event was sponsored by leading tech organizations, including AWS, GitHub, Convex, and LaunchDarkly.  As an official LaunchDarkly partner,  Terazo unlocks LaunchDarkly’s extensive feature flagging, experimentation, and targeting features with processes and training to optimize any team’s use of this powerful toolset.  Terazo amplifies LaunchDarkly’s capabilities through our Velocity solution, which delivers a customized, accelerated onboarding process.  When LaunchDarkly asked Terazo to work on a concept linking Amazon Bedrock AI Agents, LaunchDarkly, and Terazo’s services, we saw that as a great focus for the hackathon.  

Terazo developed and entered DeeDee, an innovative AI assistant designed to streamline project delivery.  The DeeDee AI assistant won a top-five spot out of 30 participating teams, underlining the exciting future of artificial intelligence solutions.  By creating an AI self-service business analysis assistant in 1 day, Terazo proved its capability to create innovative solutions with AI for customers alongside LaunchDarkly and other partners.

Meet DeeDee: AI assistance for project delivery.

Terazo is excited to announce the development of DeeDee, an innovative AI assistant made to streamline major product delivery steps. DeeDee will empower delivery directors, product owners, solution architects, and other business roles. DeeDee offers intelligent assistance and support for business analysis, architectural design formulation, and user story backlog generation processes.

DeeDee is a collection of five specialized AI agents:

  • Dee Dee, a delivery director
  • Poli, a senior product owner
  • Parsy, a junior product owner
  • Archie, a solution architect specializing in AWS Cloud services
  • Ozzy, a solution architect specializing in Open Source services

Deedee can switch between AI agents whenever generated content fails user validation, creating an adaptive user experience.

How does DeeDee work?

These five AI agents comprise DeeDee, which incorporates a “human-in-the-loop” step for validation.  DeeDee is powered by LaunchDarkly’s Quick Starts and GitHub’s Visual Studio Code Copilot extension, enabling rapid prototyping and efficient collaboration between team members and departments. We built DeeDee’s user-friendly interface with Next.js to improve the experience of its internal users.

DeeDee joins Terazo’s other AI-based solutions, which have been custom-built and put into production for clients.  Another recent innovation leveraged multi-agentic AI to generate custom and compliant SMS messages.  This solution generated a 30% increase in lead re-engagement.  In 2024, Terazo has made significant strides in AI solutions, leveraging our existing and trusted digital engineering project delivery.  We combine AI technologies with API engineering and agile development practices, further accelerating the fast production our clients expect.

Stay tuned for updates as we continue to refine and expand DeeDee’s capabilities, empowering our teams to deliver exceptional solutions for our clients. Schedule your no-risk consultation with Terazo today to learn how you can put emerging AI technologies, including DeeDee, to work powering your customer engagement and retention strategy.

Watch a demo of Dee Dee below

Torie Flood

Share on Social Media

Related Posts

Real Estate

PrimeStreet Triples Response Rates By Integrating Twilio Flex and AI

Blog

The Science Behind Terazo’s Top-Five Finish at Summer of Galaxy’s Hackathon

Blog

How can a large language model (LLM) improve customer issue resolution?

Blog

Integrated Customer Engagement: Terazo & Twilio Improve Customer Experiences at Scale