Inference Performance Engineer - Benchmarking & Capacity

Bitdeer Technologies Group; Singapore; On-site; SGD 120,000 - 170,000; Full time

Benefits offered by this job

Welfare benefits

Training and mentoring

Job summary

Bitdeer Technologies Group in Singapore seeks an Inference Performance Engineer to align MaaS benchmarks with real production workloads. You will characterize traffic shapes, keep benchmarking honest, and convert results into capacity plans and unit-cost economics for the business.

The role requires a data-first mindset, experience with GPU-backed inference workloads, and collaboration with infrastructure and finance to ensure metrics and costs reflect reality.

Qualifications

• 5+ years of experience in performance engineering or applied ML systems on large-scale, latency-sensitive distributed services.
• Experience characterizing production traffic statistically — distributions of request rate, prompt/output lengths, concurrency, burstiness.
• Knowledge of modern LLM inference internals and serving frameworks.
• Experience designing, running, or auditing load-testing and benchmarking tooling and reconciling output with live telemetry.
• Fluency with observability stacks and metric definitions used across engineering, product, and leadership dashboards.

Responsibilities

• Characterize production traffic and build models of prompt/output lengths, concurrency, and burstiness.
• Own validation loop between load-testing and live production behavior; update benchmarks as production shifts.
• Translate workload data into capacity forecasts and cost models for GPU utilization and pricing.
• Define canonical metric definitions (latency percentiles, throughput, utilization) across teams.
• Build dashboards and reports to give leadership a trusted view of performance and cost trends.

Skills

Performance engineering

Capacity planning

Applied ML systems

LLM inference workloads

Tools

Python/Pandas

SQL

Spark

k6

Locust

Job description

Bitdeer Technologies Group in Singapore seeks an Inference Performance Engineer to align MaaS benchmarks with real production workloads. You will characterize traffic shapes, keep benchmarking honest, and convert results into capacity plans and unit-cost economics for the business.

The role requires a data-first mindset, experience with GPU-backed inference workloads, and collaboration with infrastructure and finance to ensure metrics and costs reflect reality.