AI Engineer

Job Description

About Scrumconnect Consulting:

Scrumconnect Consulting is a multi-award-winning digital consultancy, recognised for delivering impactful technology solutions across UK government departments. Our work has positively influenced the lives of over 40 million UK citizens. With a strong commitment to user-centred design and agile delivery, and more to deliver innovative digital services that matter

Overview

Judicial Transcribe is an AI-enabled platform designed to deliver automated transcription and AI-assisted processing of judicial audio recordings. The service is already deployed at scale and is now entering the next phase of operational maturity.

The current focus is to stabilise the live service, strengthen engineering resilience, and extend capabilities including transcription from uploaded audio files, anonymisation/redaction, and AI-driven summarisation.

The programme also acts as a

testbed for evolving AI engineering practices

, helping to establish reusable patterns for future AI-enabled products and services.

To support this phase, the team is seeking

two experienced AI Engineers

for an initial six-month engagement.

Role Focus

The AI Engineer will focus on

AI capability development and production maturity

, helping to improve reliability, scalability, and engineering practices across the service.

You will work closely with engineers, product teams, and delivery leads to expand core functionality while ensuring robust, secure, and well-governed AI workflows.

Key Responsibilities

AI Service Development

• Support and enhance the live Transcribe service, focusing on AI components.
• Design and implement transcription workflows for uploaded audio files.
• Develop anonymisation and redaction capabilities aligned with publication requirements.

AI Integration \& Capabilities

• Support the transcript summarisation proof of value (PoV).
• Enable reuse of existing summarisation capabilities through API-based integration.
• Work with internal API platforms to expose and consume AI capabilities.

AI Engineering Practices

• Strengthen AI pipeline patterns including:

+ Extract → Chunk → Prompt → Evaluate workflows

• Integrate evaluation mechanisms and Responsible AI controls into AI features.
• Contribute to the safe and structured use of AI coding assistants within the development lifecycle.

DevSecOps \& Operational Maturity

• Improve AI-specific CI/CD practices, testing frameworks, and deployment processes.
• Enhance observability, monitoring, and production readiness for AI services.
• Support improvements in system resilience and operational stability.

Knowledge Sharing \& Engineering Resilience

• Reduce single-point-of-failure risks by documenting systems and sharing knowledge.
• Contribute to engineering standards, documentation, and operational playbooks.

Expected Outcomes

During The Six-month Engagement The Team Aims To Deliver

• Stable and sustainable live support capability for the Transcribe platform.
• Delivery of audio file transcription capability.
• Delivery or controlled rollout of anonymisation/redaction features.
• Clear integration pathway for transcript summarisation services via API.
• Improved deployment confidence and release cadence.
• Reduced dependency on individual engineers through improved documentation and shared knowledge.
• Strengthened AI engineering practices and operational maturity across the service

Required Skills \& Experience

Core Technical Skills

• Strong experience building AI-enabled applications in production environments.
• Experience with speech-to-text / transcription pipelines or similar AI workloads.
• Hands-on experience with LLMs, prompt engineering, and evaluation workflows.
• Experience designing AI pipelines and data processing workflows.

Engineering \& DevOps

• Experience with CI/CD pipelines and DevSecOps practices.
• Knowledge of API integration and microservice architectures.
• Experience implementing observability and monitoring for AI systems.

Additional Experience

• Familiarity with

check(event) ; career-website-detail-template-2 \=\> apply(record.id,meta)" mousedown\="lyte-button \=\> check(event)" final-style\="background-color:#68B54C;border-color:#68B54C;color:white;" final-class\="lyte-button lyteBackgroundColorBtn lyteSuccess" lyte-rendered\=""\>