AI Benchmark Engineer Intern: Build GroundTruth Evaluations

NUS Enterprise; Singapore; On-site; SGD 60,000 - 100,000; Full time

Job summary

NUSX Labs is the technical arm of NUSX, turning research into inventions, patents, and licensed deals. We are a hands-on team where every AI system shipped is graded against a ground-truth dataset.

You will build, evaluate, and improve the data pipeline and benchmarking framework under direct supervision of an engineer. Join a small team with real stakeholder-facing projects and deployment experience, applying Python, Claude, and Antigravity to develop high-quality data tools for patent-related

Qualifications

• Proven ability to build and ship a data pipeline.
• Familiarity with evaluation concepts (precision, recall, accuracy) and LLM evaluation metrics.
• Willingness to learn domain knowledge (e.g., patents).
• Basic familiarity with version control (Git).
• Willingness to use AI tools (Claude, Antigravity) daily.
• Nice to have: basic SQL; cloud familiarity (GCP/AWS); understanding data engineering concepts; comfortable reading/writing in English and Chinese.

Responsibilities

• Build benchmark datasets across multiple document types using Claude and Antigravity, drafting and reviewing with an annotation tool.
• Run the blind-test evaluation matrix on vendor deliverables and compare against ground truth.
• Identify edge cases, structural mismatches, and hallucinations and document them.
• Review system interfaces to determine data storage for screens and build gold field and schema catalogs.
• Compile findings into quality-acceptance reports for team sign-off on deliveries.
• Scope may extend beyond listed responsibilities.

Tools

Python

Git

Claude

Antigravity

SQL

Job description

NUSX Labs is the technical arm of NUSX, turning research into inventions, patents, and licensed deals. We are a hands-on team where every AI system shipped is graded against a ground-truth dataset.

You will build, evaluate, and improve the data pipeline and benchmarking framework under direct supervision of an engineer. Join a small team with real stakeholder-facing projects and deployment experience, applying Python, Claude, and Antigravity to develop high-quality data tools for patent-related