AI Software Engineer

Coltech; Greater London; Hybrid; GBP 90,000 - 130,000; Full time

Job summary

Coltech in London is seeking an experienced AI Software Engineer to design and implement intelligent coding agents and AI-driven workflows across the SDLC, with a focus on the GitHub ecosystem and agent-based architectures.

This hybrid, long-term role requires hands-on experience with Vibe Coding, code tooling, and CI/CD integration, plus the ability to document, test, and govern reusable Skills libraries and PromptOps standards.

Qualifications

• Recent commercial experience with Vibe Coding and AI-assisted software development.
• Hands-on experience with the GitHub ecosystem and complex GitHub-based automation.
• Experience with AI coding platforms such as Claude Code, Codex or GitHub Copilot.
• Ability to design agent-based architectures and multi-step AI workflows.

Responsibilities

• Design, build and maintain advanced AI coding and automation agents within GitHub-based workflows.
• Develop multi-step agentic workflows that extend beyond basic prompting.
• Tune and refine agents to improve reliability, performance and output quality.
• Build robust evaluation frameworks for testing AI-generated code, prompts, Skills and automated workflows.
• Establish human-in-the-loop review and approval processes before solutions are shared or deployed.
• Create and govern reusable prompt libraries, instruction sets and organisational PromptOps standards.
• Develop comprehensive Skills libraries that can be adopted across multiple engineering teams.
• Integrate AI agents with CI/CD pipelines, GitHub workflows and MCP servers.
• Embed AI-assisted development practices across the full SDLC.
• Work closely with engineering, architecture, security and platform teams.
• Maintain high standards of coding, testing, documentation and software governance.

Skills

Python

TypeScript

Rust

Golang

Java

LLM observability

CI/CD

Tools

Claude Code

Codex

GitHub Copilot

Model Context Protocol

Job description

AI Software Engineer – Vibe Coding & Agentic Development

Location: London – hybrid, 2 days per week onsite

Engagement: Long-term contract

IR35 Status: Inside IR35

Role Overview

We are seeking an experienced Software Engineer with recent, hands‑on Vibe Coding and agentic development experience to design and implement intelligent solutions across the Software Development Life Cycle.

The role will focus on the GitHub ecosystem, building specialised coding agents, custom agents, reusable Skills libraries and AI-driven workflows. You will ensure these solutions meet rigorous engineering, security and quality standards, with appropriate human review before they are shared or deployed.

Key Responsibilities

• Design, build and maintain advanced AI coding and automation agents within GitHub-based workflows.
• Develop multi-step agentic workflows that extend beyond basic prompting.
• Tune and refine agents to improve reliability, performance and output quality.
• Build robust evaluation frameworks for testing AI-generated code, prompts, Skills and automated workflows.
• Establish human-in-the-loop review and approval processes before solutions are shared or deployed.
• Create and govern reusable prompt libraries, instruction sets and organisational PromptOps standards.
• Develop comprehensive Skills libraries that can be adopted across multiple engineering teams.
• Integrate AI agents with CI/CD pipelines, GitHub workflows and Model Context Protocol servers.
• Embed AI-assisted development practices across the full SDLC.
• Work closely with engineering, architecture, security and platform teams.
• Maintain high standards of coding, testing, documentation and software governance.

Essential Skills and Experience

• Recent commercial experience with Vibe Coding and AI-assisted software development.
• Strong hands‑on experience with the GitHub ecosystem and complex GitHub-based automation.
• Practical experience using tools such as Claude Code, Codex, GitHub Copilot or similar AI coding platforms.
• Experience building custom coding agents, specialised agents and reusable Skills.
• Strong programming expertise in either Python or TypeScript.
• Solid understanding of agent-based architectures and multi-step AI workflows.
• Experience integrating AI solutions with CI/CD pipelines.
• Knowledge of Model Context Protocol and MCP server integration.
• Experience developing evaluation frameworks for AI-generated outputs.
• Ability to write clear, reusable prompts, instructions and technical documentation.
• Strong understanding of software engineering standards, testing and code quality.
• Professional development experience with Rust, Golang or Java.
• Experience establishing PromptOps standards across engineering teams.
• Knowledge of LLM observability, guardrails and AI governance.
• Experience implementing human-in-the-loop approval processes.
• Previous experience leading AI-driven engineering or developer productivity initiatives.