Principal AI Engineer: Agentic Systems & Data Quality

Intellias; Poland; On-site; PLN 180,000 - 260,000; Full time

Job summary

Intellias in Poland is seeking an engineering professional to build production-grade agentic and LLM systems. You will design workflows across PDFs, transcripts, files, and databases, surfacing confidence, gaps and provenance to end users.

You will implement ingestion pipelines, quality checks, evaluation guardrails, dashboards, and monitoring, working with platform engineering, data sourcing, and PMs to translate business expectations into measurable, automated standards.

Qualifications

• Proven experience building production agentic and LLM systems across multiple sources (PDFs, transcripts, files, databases).
• Hands-on experience engineering document ingestion and extraction pipelines with robust quality controls.

Responsibilities

• Build agentic workflows that reason over research content and present gaps and confidence to end users.
• Engineer automated quality checks on unstructured source content before ingestion.
• Build vendor delivery validation to detect defects and feed improvements back to sourcing teams.
• Develop evaluation guardrails for groundedness, provenance and citation of outputs.
• Ship monitoring dashboards showing data-quality findings and coverage gaps for engineers and PMs.
• Collaborate with platform engineering, data sourcing and portfolio managers to define measurable quality standards.

Skills

Python

LLM workflows

Production systems

Data ingestion

CI/CD

Tools

Snowflake

Linux/UNIX

Git

Jira

Job description

Intellias in Poland is seeking an engineering professional to build production-grade agentic and LLM systems. You will design workflows across PDFs, transcripts, files, and databases, surfacing confidence, gaps and provenance to end users.

You will implement ingestion pipelines, quality checks, evaluation guardrails, dashboards, and monitoring, working with platform engineering, data sourcing, and PMs to translate business expectations into measurable, automated standards.