Adaptive Model Routing & Agent Evaluation Scientist

Bitdeer Technologies Group; Singapore; On-site; SGD 180,000 - 260,000; Full time

Benefits offered by this job

Training program

Mentoring & growth opportunities

Competitive benefits

Job summary

Bitdeer Technologies Group is seeking a highly skilled AI Engineer to lead the evaluation and orchestration of agentic inference. The role focuses on building robust LLM evaluation pipelines, designing adaptive routing strategies, and ensuring reliable deployment within MaaS and platform teams.

You will work on cutting-edge research, develop practical prototypes, and contribute to the scale of Bitdeer AI Lab's intelligent services, with emphasis on performance, cost, and latency trade-offs in

Qualifications

• Hands-on experience in LLM evaluation, agentic systems, applied machine learning, or adaptive inference.
• Ability to design or operate evaluation pipelines for LLMs or agents, with task-level and trajectory-level metrics.
• Experience with multi-model APIs, agent harnesses, traffic replay, shadow evaluation, A/B testing, or production model monitoring is highly preferred.

Responsibilities

• Own and extend LLM and agent evaluation pipeline; develop representative task suites.
• Research and prototype adaptive model-routing strategies inside a shared agent harness.
• Collaborate with MaaS and platform teams on production integration and reliability.

Skills

LLM evaluation

Agentic systems

Applied ML

Adaptive inference

Education

BSc in CS

MSc in ML

PhD in EE

Tools

Python

PyTorch

Evaluation tooling

Job description

Bitdeer Technologies Group is seeking a highly skilled AI Engineer to lead the evaluation and orchestration of agentic inference. The role focuses on building robust LLM evaluation pipelines, designing adaptive routing strategies, and ensuring reliable deployment within MaaS and platform teams.

You will work on cutting-edge research, develop practical prototypes, and contribute to the scale of Bitdeer AI Lab's intelligent services, with emphasis on performance, cost, and latency trade-offs in