/projects$cat xenia-martech.html

Xenia · AI-Driven Omni-Channel Martech

iLabs · 2025 · architected & built the agentic layer

• LangChain
• deep agents
• LangGraph
• GA4
• BigQuery
• Crawl4AI
• Firecrawl

the problem

Marketing across web, mobile, email and SMS is a coordination nightmare: pull the data, understand the audience, decide the journey, then create and deliver content per channel. Xenia's goal was to make that autonomous — a platform that plans and runs campaigns.

what i built

I architected and developed the agentic AI layer. Using LangChain deep agents, I built complex, graph-based user journeys, and a server architecture that manages multiple concurrent orchestrator agents — each directing specialized sub-agents with targeted tools:

• Database querying — pulling the platform's own structured data.
• Marketing intelligence — GA4 and BigQuery for real audience signals.
• Autonomous web scraping — Crawl4AI and Firecrawl for live external context.
• Content creation — dedicated push & pull channel agents that generate and deliver targeted material across web, mobile, email and SMS.

The main engineering challenge was concurrency — running many orchestrators, each with its own sub-agents, reliably and at the same time.