Deine Aufgaben
Phase 1: Transform our product into a fully agentic system
What You Will Actually Do
- Redesign Rocket (MiceDesks core product) into a multi-agent architecture with autonomous decision-making and workflow execution.
- Build agents that automate all major MICE processes: availability checks, pricing logic, offer creation, approvals, contracting, follow-ups, and post-booking operations.
- Implement robust LLM pipelines using:
- AWS Bedrock
- Local & fine-tuned LLMs
- Vector DBs & embedding strategy
- Knowledge Models (Hotel Specific Knowledge HSN)
- Create deterministic, reliable agent behaviors in a complex operational environment.
- Architect and scale the global hospitality data lake, unifying PMS, CRM, RFP, event, and operational datasets.
- Create a canonical data model that scales from single properties clusters global chains.
- Build the intelligence layer powering:
- Predictive demand
- Revenue optimization
- Process automation
- Profile graph models
- Operational insight engines
What You Will Actually Do
- Build autonomous multi-agent workflows inside Rocket Own the full LLM engineering stack,
- Design the loop between PMS data Data Lake Agents UI,
- Develop fine-tuned models for extraction, reasoning, pricing, and anomaly detection,
- Drive architectural decisions with founders and engineers,
- Shape the intelligence layer that will scale across thousands of hotels.