Inference Performance Engineer - MaaS

About the role

Bitdeer, a technology company for Bitcoin mining and AI cloud, is looking for an Inference Performance Engineer for its MaaS platform. The role closes the gap between outdated load-test assumptions and real production traffic, characterizing actual request patterns, validating benchmarks against live telemetry, and turning the results into capacity and cost-per-token models that the business can plan around.

Key responsibilities

• Profile real production request streams across chat, completions, embeddings and fine-tuning endpoints, and build statistical models of prompt/output length, concurrency and burstiness by model and customer segment
• Validate load-testing and benchmarking harnesses against production traffic, flagging drift before it affects capacity or optimization decisions
• Turn workload characterization and GPU utilization data into capacity forecasts and cost-per-token or cost-per-request models, working with infrastructure and finance teams
• Define and maintain canonical metric definitions (latency percentiles, throughput, utilization) and document instrumentation requirements and aggregation windows
• Build dashboards and reports that give engineering, product and leadership a shared view of performance and cost trends, and act as reviewer for performance and cost claims

Required profile

• 5+ years in performance engineering, capacity planning or applied ML systems work on large-scale, GPU-backed services
• Experience characterizing production traffic statistically, using Python/pandas, SQL, Spark or a comparable stack
• Experience designing, running or auditing load-testing and benchmarking tooling (k6, Locust, custom generators, MLPerf-style suites) and reconciling output with live telemetry
• Comfortable partnering with FinOps and infrastructure teams on GPU utilization and cost modeling

Required skills

• Working knowledge of LLM inference internals (continuous batching, KV-cache management, prefill/decode scheduling) and serving frameworks such as vLLM, TensorRT-LLM, SGLang or Triton Inference Server
• Fluency with observability stacks such as Prometheus, Grafana and distributed tracing tools

What we offer

• A culture that values authenticity and diversity, with an inclusive environment and open workspaces
• Fast-growing company with direct impact on the digital asset industry
• Personal accountability, autonomy, fast growth and learning opportunities
• Welfare benefits and development opportunities such as training and mentoring