Research Engineer: AI Data Pipelines & Evaluation (Singapore)

Clera; Singapore; On-site; SGD 128,000 - 218,000; Full time

Benefits offered by this job

Visa sponsorship available

Job summary

Clera on-site in Singapore is seeking an engineer to join an AI evaluation and reinforcement learning team. You will transform real-world workflows into scalable, synthetic training tasks and devise metrics to measure improvements in model capabilities.

You will build end-to-end generation pipelines, work with domain SMEs, and develop tools to mutate, validate, and enhance synthetic tasks while analyzing agent performance to identify what helps learning and where gaps remain.

Qualifications

• Two or more years of software engineering, ML engineering, or AI research experience with data pipelines or ML infrastructure.
• Hands-on experience applying synthetic data methods to end-to-end generation pipelines.
• Proficiency in Python and Linux environments with Docker.
• Understanding of synthetic data quality, evaluation metrics, and data limitations.
• Experience building evaluation benchmarks, frameworks, or testing environments for AI agents or language models.

Responsibilities

• Build pipelines that transform workflows into synthetic training tasks.
• Collaborate with subject-matter experts to create tasks across domains.
• Develop methods and tools to generate, mutate, validate, and improve synthetic tasks.
• Analyze agent performance to learn what tasks teach and where they fail.
• Create metrics for task diversity, realism, learnability, and quality.

Skills

Python

Linux

Docker

Data pipelines

ML infrastructure

Evaluation frameworks

Job description

Clera on-site in Singapore is seeking an engineer to join an AI evaluation and reinforcement learning team. You will transform real-world workflows into scalable, synthetic training tasks and devise metrics to measure improvements in model capabilities.

You will build end-to-end generation pipelines, work with domain SMEs, and develop tools to mutate, validate, and enhance synthetic tasks while analyzing agent performance to identify what helps learning and where gaps remain.