Position Summary
We are seeking an expert GenAI engineer with 8+ years of relevant software, data or AI engineering experience to build and operate production AI applications for a Saudi organization, using n8n to connect AI features to existing services. The work must be reliable, testable and well documented. Only client-approved models, tools and hosting may be used, and client data must never go to public AI services or be used for training outside approved scope.
Responsibilities
- Build and customize LLM applications aligned with approved use cases and architecture.
- Implement production n8n workflows connecting AI to authenticated APIs, enterprise services and approved data.
- Design for retries, timeouts, duplicate events, failed steps and controlled manual recovery.
- Write supporting code, integrations and data transformations where n8n alone falls short.
- Set up prompt/config versioning, repeatable evaluation, output validation and logging; investigate issues with QA and DevOps.
- Review the mid-level engineer's work and deliver source, documentation, operating guidance and knowledge transfer.
Requirements
Essential
- 8+ years in software, data or AI engineering, including recent production GenAI work.
- Current, separately demonstrable capability in both GenAI and n8n (8 years in each tool is not required).
- Relevant bachelor's degree (CS, AI, data science, software engineering) or equivalent.
- Proven production LLM/GenAI delivery, beyond experiments or prompt demos.
- Hands-on n8n: webhooks/APIs, authentication, secrets, error/retry branches, deployment and support.
- Strong Python (or comparable) skills and ability to review integration code.
- Able to evaluate outputs against business needs, control data access and explain reliability/security trade-offs.
- Compliance with client-approved models, tools and hosting.
Preferred
- RAG or document-data integration
- Arabic-language applications, government/regulated services, AI quality testing
- CI/CD for AI/workflow assets, containerization, model/provider portability
