Lead AI Systems Engineer

VECTRA International BV; Polska; Hybrid; PLN 420,000 - 660,000; Full time

Job summary

VECTRA International BV seeks an experienced backend engineer to lead the design, deployment, and security of an AI-powered assurance suite. You will own architecture, prompt engineering, and private RAG pipelines in private EU endpoints.

This role requires 6+ years in backend software engineering with 3+ years in production LLM/RAG systems, plus GDPR compliance knowledge, FastAPI, Docker, and PostgreSQL with pgvector to safeguard data.

Qualifications

• 6+ years in backend software engineering with 3+ years in production LLM/RAG system deployment and infrastructure security.
• Tech Stack: Python, LangChain/LlamaIndex, vLLM/Ollama, Private Cloud EU endpoints, FastAPI, PostgreSQL (pgvector), Docker.
• Compliance & Security: Working knowledge of GDPR data security mandates, tenant data isolation, and API encryption standards.

Responsibilities

• Direct the clean-slate design, LLM orchestration, and deployment of the AI-powered assurance suite.
• Own system architecture, prompt engineering, private RAG pipelines, local/private LLM hosting, and the proprietary multi-stage concierge isolation layer that air-gaps client data from public models.
• Architect private, locally hosted model environments to satisfy data privacy requirements.
• Build and enforce a three-tier architecture to guarantee zero proprietary data leakage to public models.
• Purge legacy training configurations and re-train models on client compliance standards.
• Lead technical setup and data mapping for assurance workflows and audit modules.
• Develop confidence-scoring thresholds and QA triggers routing ambiguous sections to humans.

Skills

Backend engineering

LLM deployment

Data security

Tools

Python

LangChain/LlamaIndex

vLLM/Ollama

FastAPI

PostgreSQL (pgvector)

Docker

Private Cloud EU endpoints

Job description

Direct the clean-slate design, LLM orchestration, and deployment of the AI-powered assurance suite. This role owns system architecture, prompt engineering, private RAG pipelines, local/private LLM hosting, and the proprietary multi-stage concierge isolation layer that air-gaps client data from external public LLMs.

2. Key Responsibilities

Clean-Slate RAG Architecture: Deploy a production-grade Retrieval-Augmented Generation pipeline using vector databases and semantic search to ingest and benchmark client policies.

Private & Locally Hosted Model Deployment: Architect private, locally hosted model environments (e.g., vLLM, Llama, open-weights Mistral/Llama models, or private Azure/AWS EU enterprise endpoints) to satisfy strict data privacy requirements that prohibit public OpenAI/Claude ingestion.

Three-Stage Concierge Isolation Layer: Build and enforce the three-tier architecture (Core Internal Application Concierge Application Layer External Public Data Gatherer) to guarantee zero proprietary data leakage to public models.

Model Training on Client Standards: Purge legacy training configurations and re-train models specifically on client compliance standards and site assessment templates.

Assurance Workflow Specialist Function: Lead technical setup and data mapping for specialized assurance workflows and audit compliance modules.

Human-in-the-Loop QA Protocols: Build confidence-scoring thresholds and automated flag triggers that route ambiguous assessment sections to human experts for verification.

Corporate IP Enforcement: Ensure all code, configurations, and API keys reside strictly within corporate GitHub repositories and enterprise accounts.

Candidate Profile & Qualifications

Experience: 6+ years in backend software engineering with 3+ years in production LLM/RAG system deployment and infrastructure security.

Tech Stack: Python, LangChain/LlamaIndex, vLLM/Ollama, Private Cloud EU endpoints, FastAPI, PostgreSQL (pgvector), Docker.

Compliance & Security: Working knowledge of GDPR data security mandates, tenant data isolation, and API encryption standards.