AI/ML Engineer

Insight International (UK) Ltd; Manchester; Hybrid; GBP 90,000 - 130,000; Full time

Job summary

Insight International (UK) Ltd in Manchester is seeking an experienced AI/ML engineer to design, build and run production-grade AI/ML solutions across the full lifecycle. The role focuses on Generative AI, Agentic AI, MLOps, scalable model serving and secure cloud infrastructure.

You will develop embeddings, retrieval, and grounding techniques, deploy via Vertex AI/KServe, and establish CI/CD, monitoring and governance across environments.

Qualifications

• Strong production Python engineering, API development, automated testing and distributed service patterns.
• Hands-on PyTorch experience and transformer understanding.
• Experience with ReAct or comparable agent patterns, including tool calling and memory.
• Production RAG, embeddings, indexing and grounding techniques.
• GCP Vertex AI and cloud compute/storage experience; GKE advantageous.
• Docker packaging and serving via Kubernetes-native platforms.
• MLOps lifecycle, autoscaling and continuous delivery practices.
• Delivery in regulated enterprise with security, privacy and risk controls.
• Technical leadership, architecture reviews and cross-functional collaboration.

Responsibilities

• Design enterprise AI apps using LLMs, transformer architectures and RAG.
• Build agents with tools, memory, reasoning and workflow orchestration.
• Develop and evaluate ML solutions from experimentation to production.
• Create embedding pipelines, semantic retrieval and grounding for responses.
• Package models with Docker and deploy via KServe or Vertex endpoints.
• Implement CI/CD, continuous training and monitoring workflows.
• Manage model/version registries and governance-aligned release processes.
• Monitor model/data quality, drift, latency and system health.
• Optimize inference with autoscaling, GPU scheduling and cloud-native architecture.
• Enable safe progressive delivery with canary/shadow/blue-green deployments.

Skills

Python engineering

API development

Automated testing

Asynchronous services

Distributed systems

PyTorch

Agentic AI

RAG & retrieval

MLOps

Canary/shadow deployments

Cross-functional collaboration

Tools

Docker

Kubernetes

Vertex AI

GKE

LangChain

LangGraph

LlamaIndex

Prometheus

Grafana

Job description

Manchester, UK | Hybrid

ROLE PURPOSE

Design, build and operate production-grade AI and machine learning solutions across the full lifecycle. The role combines Generative AI and Agentic AI engineering with MLOps, scalable model serving, cloud infrastructure, production monitoring and responsible engineering controls.

Key responsibilities

• Design enterprise AI applications using Large Language Models, transformer architectures, Retrieval-Augmented Generation (RAG) and Agentic AI patterns such as ReAct.
• Build agents that use tools, reasoning, memory and workflow orchestration; integrate AI capabilities with enterprise platforms through secure APIs and microservices.
• Develop, evaluate and optimise machine learning and deep learning solutions using Python and PyTorch, taking work from experimentation through production.
• Create embedding, indexing, semantic retrieval and ranking pipelines for grounded AI responses and enterprise knowledge use cases.
• Package models using Docker and deploy through KServe, Vertex AI endpoints and Kubernetes-based serving platforms.
• Build CI/CD, continuous training and continuous monitoring workflows, including model evaluation and controlled promotion across environments.
• Operate feature and model registries, model versioning and reproducible release processes aligned to governance and risk controls.
• Monitor model quality, data quality, drift, latency, fairness signals, infrastructure health and service-level objectives.
• Optimise inference performance and cost through autoscaling, GPU scheduling, resource management and cloud-native architecture.
• Enable safe progressive delivery using canary, shadow and blue/green deployments, rollback controls and A/B testing.
• Collaborate with Product, Data Science, Platform, Architecture, Security and Risk teams; establish reusable patterns and engineering standards.

Mandatory skills and experience

Capability

Required experience

Core engineering

Strong production Python engineering, API development, automated testing and asynchronous or distributed service patterns.

Deep learning

Hands-on PyTorch experience and strong understanding of transformer architecture, inference and model evaluation.

Agentic AI

Experience with ReAct or comparable agent patterns, including tool calling, memory, reasoning and orchestration.

RAG & retrieval

Production RAG, embeddings, chunking, indexing, vector search, retrieval/ranking and grounding techniques.

Google Cloud

GCP knowledge with Vertex AI, cloud compute/storage and cloud databases; GKE experience is advantageous.

Containers & serving

Docker plus model packaging and serving using KServe, Vertex AI endpoints or equivalent Kubernetes-native platforms.

MLOps lifecycle

Scale & release

Cost-efficient autoscaling, GPU scheduling, canary and shadow deployment, rollback strategies and A/B testing.

Other experience

• LangChain, LangGraph, LlamaIndex or comparable orchestration frameworks.
• Kubernetes/GKE, Harness or equivalent delivery tooling, GitOps and infrastructure automation.
• Prometheus, Grafana, Dynatrace or similar observability platforms.
• Delivery in a regulated enterprise with security, privacy, model risk and responsible AI controls.
• Technical leadership, architecture reviews, mentoring and cross-functional stakeholder collaboration.

What success looks like

• AI services move safely from experiment to reliable, monitored production with repeatable delivery controls.
• Model serving is secure, resilient, scalable and cost-efficient, with measurable quality and operational performance.
• Reusable patterns improve delivery speed while supporting transparency, governance and risk management.