Leading frontend architecture and AI systems engineering for
Azara AI — JLL's enterprise conversational analytics platform.
Conversational AI & Agent Systems
- Architected the conversational UI layer for Azara AI — a chat-driven analytics
assistant rendering AI-generated markdown, dynamic tables, and Highcharts
visualizations from real-time streaming responses.
- Designed a LangGraph-powered multi-agent pipeline on Databricks Genie — coordinating
agents for rephrasing, visualization selection, and reasoning, with shared state,
human-in-the-loop checkpoints, and resumable execution for production reliability.
- Built the streaming UI layer (Vercel AI SDK) using token-level streaming and
optimistic UI patterns — reducing perceived latency and making responses feel
conversational, not batch-processed.
Platform Engineering & Developer Tooling
- Built an internal MFE CLI tool to bootstrap apps with pre-configured best practices
(Jest, ESLint, CodeClimate, commit-lint) — cutting team onboarding time by 80% across 10+
teams.
- Owned delivery of three core MFE modules (MyAnalytics, Dashboard Configuration,
Subscribe to Dashboard) via Webpack Module Federation — serving 500+ enterprise users.
- Delivered full-stack features across React + TypeScript, NestJS, PostgreSQL, Redis,
and GraphQL — owning UI to API, data modeling, and performance optimization.
Performance, DevOps & Engineering Culture
- Migrated 25+ GitHub repos to modern CI/CD
pipelines with automated release notes and semantic versioning — reducing release
overhead by 30%.
- Resolved LCP regression (10s → 5s) via render
profiling, code-split redesign, and API waterfall elimination.
- Mentored junior engineers via PR reviews and guild sessions on performance, MFE
governance, and AI-integrated UI — influencing platform standards beyond team scope.
React.js
TypeScript
NestJS
PostgreSQL
Redis
GraphQL
LangGraph
Databricks Genie
Vercel AI SDK
Webpack Module Federation
Highcharts
Docker
Jest
Storybook