Azure Observability - AI Monitoring - Contract

London Source: WhatJobs

£600 - £650 Outside IR35

Azure Observability / Grafana Engineer - AI Monitoring - Contract

Method Resourcing are supporting a major organisation as they expand their use of AI and intelligent automation across the business.

As the number of AI applications and agents increases, the organisation wants to establish a central monitoring and observability capability that provides a clear view of how AI is being used, by whom, at what cost, and how effectively it is performing.

We are looking for an experienced Azure Observability / Monitoring Engineer to initially design the monitoring solution before supporting its build and implementation.

The Project

The engagement will begin with an initial 4-week design and discovery phase, before moving to build.

You will work with technical and business stakeholders to understand the existing Azure and AI landscape, identify the available telemetry and data sources, and establish what needs to be captured to provide meaningful oversight of AI consumption.

The solution will be built within Microsoft Azure, with Grafana used to provide dashboards, monitoring and visualisation.

Once the design is agreed, the project will move into the build and implementation phase.

What will be monitored?

The monitoring capability will provide visibility across the organisation's AI estate, potentially including:

AI usage and traffic - requests, interactions, volumes and usage patterns across different AI applications and agents
Cost and consumption- tracking AI expenditure and identifying which applications, teams or services are driving cost
User and application activity - understanding who consuming AI services is, when they are being used and through which applications
Model usage - visibility of which AI models and services are being consumed and their relative utilisation
Token consumption - monitoring input/output token volumes and associated costs where applicable
Performance- response times, latency, throughput and availability
Reliability - failed requests, exceptions, timeouts, retries and service degradation
Agent performance - monitoring individual agents, workflows and dependencies to understand how effectively they are operating
Adoption and trends - identifying changes in usage over time and areas of high or low adoption
Capacity and scalability - understanding peaks in demand and potential constraints as AI usage increases
Cost optimisation - identifying inefficient usage patterns and opportunities to reduce unnecessary consumption
Business reporting / ROI - combining technical usage and cost information with appropriate business metrics to help assess the value being delivered by AI initiatives