Back to Projects

// Freelance client work via Aqlity

IRIS Global

I built a full-stack, multi-agent AI research platform for NuqiWealth covering global stock markets across seven exchanges: NYSE, Nasdaq, NYSE American, London Stock Exchange, Cboe Europe, Dubai Financial Market, and Abu Dhabi Securities Market.

Tech Stack

ReactNext.jsPythonMulti-Agent AI ArchitectureDockerRedisRabbitMQ

How it works

Specialized agents handle fundamentals, technical charts, news sentiment, screening, and rankings, routed through a layer that decides which agents a given question actually needs.

  • Built an automatic answer-quality check that scores every response and falls back to a live web search when internal data isn't sufficient.
  • Built a background scheduler that pre-analyzes stocks overnight so results are fast the next morning.
  • Connected the platform directly to a user's real portfolio and watchlist for personalized analysis.

Why I built it

Investors tracking global markets across multiple exchanges usually need several disconnected tools, and most AI research assistants either run every agent on every question (slow) or fail silently when their data is incomplete.

How I approached it

I built a routing layer that decides which specialized agents a question actually needs instead of running all of them every time, added a quality check that falls back to live web search when internal data falls short, and pre-computed overnight analysis so the platform feels instant during market hours.

IRIS Global screenshot 1
IRIS Global screenshot 2
IRIS Global screenshot 3
IRIS Global screenshot 4
IRIS Global screenshot 5
IRIS Global screenshot 6
IRIS Global screenshot 7