AI Engineer – Data Provenance & Knowledge-Base Pipelines

Qonnectiq; Indonesia; On-site; IDR 1,073,345,000 - 2,146,691,000; Full time

Job summary

Qonnectiq seeks an AI Engineer to design and own the core database schema, enabling versioning, provenance, and auditable data across a large document archive. You will implement parsing pipelines for PDFs, ensure traceability of every figure to its source, and build a robust validation layer.

You will lead data engineering and AI efforts, including extraction of operational parameters, equipment specs, and costs, while evaluating in-house vs commercial AI options and building repeatable quality

Qualifications

• Relational DB design with historical data versioning and auditability.
• 5+ years as AI/ML engineer with hands-on data pipelines.
• Experience parsing PDFs or semi-structured formats.
• Strong technical English for cross-domain collaboration.

Responsibilities

• Design and own the database schema acting as the system's source of truth.
• Implement as-of-date/versioning logic for historical queries.
• Ensure provenance: source document, location, extraction method, confidence.
• Build PDF parsers for complex layouts and OCR-reliant documents.
• Extract structured entities like parameters, specs, readings, costs.

Skills

Relational DB design

Historical data versioning

AI/ML engineering

Data pipelines

PDF information extraction

RAG ( Retrieval-augmented generation )

Model evaluation/testing

Commercial vs in-house AI trade-offs

SQL

Graph DB / knowledge graphs

Tools

OCR tooling

Graph databases

Job description

Qonnectiq seeks an AI Engineer to design and own the core database schema, enabling versioning, provenance, and auditable data across a large document archive. You will implement parsing pipelines for PDFs, ensure traceability of every figure to its source, and build a robust validation layer.

You will lead data engineering and AI efforts, including extraction of operational parameters, equipment specs, and costs, while evaluating in-house vs commercial AI options and building repeatable quality