← Back to Projects

Zeta (AI-enabled Wealth Platform)

GenAI-assisted wealth workflows, streaming chat UI, and production-grade advisor experience.

Zeta platform thumbnail

Tech Stack

  • React / Next.js
  • GenAI Streaming
  • TypeScript
  • Java / Golang
  • Python
  • Temporal
  • Kafka
  • AWS S3

Key Contributions

  • Built GenAI chat components integrated with LLM-powered services.
  • Implemented real-time streaming API patterns for incremental AI responses.
  • Developed Java-based video training platform with signed URL authentication.
  • Added speech-to-text capabilities for voice-driven interactions.
  • Owned the frontend architecture for high-performance advisor workflows.

The Challenge

  • Handling incremental responses without duplicated text
  • Maintaining UI responsiveness under continuous updates

The Solution

  • Stable message IDs + patch-based updates
  • State partitioning and memoization strategies

Outcomes

  • More stable streaming UX and fewer transcription/chat glitches
  • Improved perceived performance and usability

Project Details

RoleFull Stack / Frontend Lead
DomainFinTech / WealthTech
Typeprofessional

Project Highlights

  • Delivered a production-grade AI wealth platform used by financial advisors.
  • Solved complex streaming UI challenges (partial updates, de-duplication).
  • Established frontend standards and led code reviews for the team.