Co-founder
CTO / Head of Operations
Sabivox LLM Support Agent
Autonomous multi-turn LLM customer-support agent cutting response time from 8 minutes to 30 seconds.
01. The Challenge
Replace slow, manual first-line customer support with an autonomous agent that can handle multi-step intent recognition reliably and forecast ticket volume.
02. The Solution
Architected end-to-end LangChain + transformer prompt chains with structured output pipelines and tool-use patterns. Deployed real-time ML inference on AWS EKS with full CI/CD (GitHub Actions), monitoring, logging, and alerting; directed 10+ engineers across backend, AI/ML, and frontend.
03. The Outcome
Cut average response time from 8 minutes to 30 seconds and achieved 94% ticket-volume forecast accuracy.
8min → 30s
Response
94%
Forecast
10+ engineers
Team