AI Inference Performance Engineer

Bitdeer Technologies Group; Singapore; On-site; SGD 180,000 - 280,000; Full time

Benefits offered by this job

Welfare benefits

Training & mentoring

Job summary

Bitdeer Technologies Group is seeking an Inference Performance Engineer to calibrate benchmarks, capacity plans, and cost projections for Bitdeer AI MaaS. You will align benchmarking with real production traffic, ensuring estimates reflect actual chat, completion, and embedding workloads.

You will collaborate with infra and finance to keep models defensible and actionable. The role emphasizes data‑driven decisions, canonical metric ownership, and close interaction with cross‑functional teams to

Qualifications

• 5+ years in performance engineering, capacity planning, or ML systems work on large‑scale services.
• Hands‑on experience characterizing production traffic statistically (distributions of request rate, token lengths, concurrency).
• Familiarity with LLM inference internals and benchmarking tooling.

Responsibilities

• Characterize real production traffic across chat, completions, embeddings, and fine‑tuning endpoints.
• Own validation loop between load‑testing, benchmarking, and live production behavior; flag drift.
• Translate workload data into capacity forecasts and cost models for GPU usage and pricing.
• Define canonical latency and throughput metrics and ensure consistent dashboards across teams.
• Build dashboards and reports that provide leadership with performance and cost insights.

Skills

Performance engineering

Capacity planning

ML systems

Python/pandas

SQL

Spark

LLM inference

Benchmarking

GPU utilization

Tools

Prometheus

Grafana

Triton Inference Server

vLLM

TensorRT-LLM

Job description

Bitdeer Technologies Group is seeking an Inference Performance Engineer to calibrate benchmarks, capacity plans, and cost projections for Bitdeer AI MaaS. You will align benchmarking with real production traffic, ensuring estimates reflect actual chat, completion, and embedding workloads.

You will collaborate with infra and finance to keep models defensible and actionable. The role emphasizes data‑driven decisions, canonical metric ownership, and close interaction with cross‑functional teams to