Remote Performance & Systems Engineer - Latency Focus

Meramia Technologies; Dubai; Remote; AED 350,000 - 620,000; Full time

Job summary

Meramia Technologies is seeking an experienced senior ML infra engineer in Dubai to optimize the end-to-end inference stack. You will map the critical path from agent loops to inference pipelines and identify where costs are highest, while shaping telemetry and profiling for the team to run.

You will own latency and cost profiling, establish baselines, and push for practical architecture changes that reduce wasted tokens and RPCs, balancing speed with compute cost and reliability.

Qualifications

• Experience with ML inference or serving platforms at scale.
• Ability to profile latency, memory, and throughput in high-throughput systems.
• Proven track record delivering measurable performance improvements.

Responsibilities

• Audit latency and cost across inference, network, and agent loop.
• Establish a profiling baseline for latency, throughput and GPU memory.
• Drive measurable reductions in p99 latency or serving cost on core workloads.
• Collaborate with teams to optimize architecture for token efficiency.

Skills

Kubernetes

GPU orchestration

Distributed services

Latency profiling

Benchmarks

Tools

CUDA kernels

Triton

vLLM

TensorRT-LLM

Job description

Meramia Technologies is seeking an experienced senior ML infra engineer in Dubai to optimize the end-to-end inference stack. You will map the critical path from agent loops to inference pipelines and identify where costs are highest, while shaping telemetry and profiling for the team to run.

You will own latency and cost profiling, establish baselines, and push for practical architecture changes that reduce wasted tokens and RPCs, balancing speed with compute cost and reliability.