Technical Lead - AI Native

Socium; Manchester; On-site; GBP 90,000 - 150,000; Full time

Job summary

Socium is building AI delivery pods where engineers and AI workers build, test, ship and run real product. The senior technical lead will own the architecture and set standards for how AI systems are built, evaluated, secured and scaled.

You’ll be the most senior technical person across the pods, owning how agents, models, tooling, data flows and security boundaries work in practice. You’ll be deciding which models to use, how agents should be orchestrated, when to build or buy, how output

Qualifications

• Hands-on experience building AI, LLM or agentic systems in production.
• Strong understanding of RAG, agent orchestration, prompts frameworks.
• Experience owning architecture decisions across security, observability and cost.
• Comfortable making technical decisions in fast-moving environments.
• Able to create standards and structure without slowing delivery.
• Strong software engineering background, ideally with .NET/C#, close to production code.

Responsibilities

• Owning the AI-native technical architecture across the pods
• Defining patterns for agents, orchestration, model selection and tooling
• Setting standards around code quality, prompt engineering, testing and evaluation
• Making technical decisions on build vs buy, re-architecture and tooling choices
• Working directly with AI workers and improving the quality of their output
• Mentoring AI Engineers and Graduates
• Working with security and compliance around data, tooling and governance
• Managing tooling subscriptions, usage costs and API spend
• Helping prove how the pod model scales beyond the first teams

Skills

Hands-on AI

Architectural ownership

Agent orchestration

Prompt engineering

C# / .NET

Security & governance

Job description

They’re building new AI delivery pods where engineers and AI workers build, test, ship and run real product.

The key thing now is making sure that doesn’t become a mess.

The business needs someone senior enough to own the technical direction, make the architecture calls and set clear standards around how AI systems are built, evaluated, secured and scaled.

You’ll be the most senior technical person across the pods, owning how agents, models, tooling, data flows and security boundaries work in practice.

The work is hands-on, but not just delivery.

You’ll be deciding which models to use, how agents should be orchestrated, when to build or buy, how output should be tested, and how to make AI workers a reliable part of the team rather than something people experiment with on the side.

What you’ll be doing

• Owning the AI-native technical architecture across the pods
• Defining patterns for agents, orchestration, model selection and tooling
• Setting standards around code quality, prompt engineering, testing and evaluation
• Making technical decisions on build vs buy, re-architecture and tooling choices
• Working directly with AI workers and improving the quality of their output
• Mentoring AI Engineers and Graduates
• Working with security and compliance around data, tooling and governance
• Managing tooling subscriptions, usage costs and API spend
• Helping prove how the pod model scales beyond the first teams

What they’re looking for

• Hands‑on experience building AI, LLM or agentic systems in production
• Strong understanding of RAG, agent orchestration, prompt engineering and evaluation frameworks
• Experience owning architecture decisions across security, observability and cost
• Someone comfortable making technical decisions in a fast‑moving environment
• Someone who can create standards and structure without slowing delivery down
• Best practice comes naturally to you and you like implementing it.
• Strong software engineering background, ideally with .NET / C#, so you can stay close to the product code and make sensible architecture decisions rather than just owning the AI layer in isolation

The bigger thing is being able to bring structure to AI delivery without turning it into heavy process.

You’ll suit this if you’ve built real AI systems, understand the detail behind the tooling, and want ownership of how small AI-native pods deliver properly.