Senior AI Platform Engineer — LLM Inference (Go/Python)

J.P. MORGAN; Scotland; On-site; GBP 110,000 - 170,000; Full time

Job summary

JPMorganChase in the United Kingdom seeks a Lead Software Engineer to advance LLM inference at scale within the AI/ML Data Platform team. You will drive performance optimization, benchmarking, and cost-efficiency across production workloads, collaborating with senior engineers and leadership to shape a high-impact platform.

You will implement quantization experiments, analyze throughput, memory usage, and per-token costs, and contribute to upgrade evaluations while ensuring secure,

Qualifications

• Formal training or certification in software engineering with Go or Python.
• Hands-on experience with LLM inference systems and production serving engines.
• Strong understanding of GPU memory architecture and quantization implications.
• Experience with quantization tradeoffs at scale and benchmarking.
• Ability to communicate technical trade-offs to peers and leadership.
• Experience operating cloud-based GPU infrastructures at scale.

Responsibilities

• Benchmark LLM workloads and establish reproducible baselines.
• Design and run quantization experiments (FP8/INT8/INT4) and measure impact.
• Support speculative decoding strategies and configuration recommendations.
• Build GPU efficiency metrics and provide data-driven optimization insights.
• Benchmark platform against external providers and industry numbers.
• Participate in inference engine upgrades and new architectures.
• Engage in GPU chaos engineering and hardware diagnostics.
• Use enterprise AI-assisted development tools and validate outputs.

Skills

Go/Python

LLM inference

GPU memory architecture

Quantization methods

Cloud GPU infra

Technical communication

Tools

vLLM

TensorRT-LLM

SGLang

LLM-D

GuideLLM

Job description

JPMorganChase in the United Kingdom seeks a Lead Software Engineer to advance LLM inference at scale within the AI/ML Data Platform team. You will drive performance optimization, benchmarking, and cost-efficiency across production workloads, collaborating with senior engineers and leadership to shape a high-impact platform.

You will implement quantization experiments, analyze throughput, memory usage, and per-token costs, and contribute to upgrade evaluations while ensuring secure,