Head of Data Science

Hamilton Barnes ?; Greater London; On-site; GBP 90,000 - 140,000; Full time

Job summary

Hamilton Barnes is looking for a Customer-Facing AI Data Scientist (LLM Architecture) to join a growing MSP practice in London. You will sit with enterprise clients, shaping LLM deployments, evaluating training and fine-tuning strategies, and ensuring commercial viability of AI solutions.

You will work at the intersection of deep technical LLM expertise and client delivery, translating complex model concepts into actionable architectures and delivery plans for named accounts.

Qualifications

• 5+ years in applied data science or ML engineering with direct exposure to LLM training or fine-tuning at a technical level.
• Hands-on experience with transformer architectures and modern ML frameworks (PyTorch, Hugging Face, or equivalent).
• Strong grounding in evaluation methodologies, RAG and fine-tuning trade-offs, and production ML deployment patterns.
• Confident and credible in front of senior technical stakeholders and able to defend design decisions in live discussions.

Responsibilities

• Act as the primary technical point of contact for enterprise clients on LLM-based engagements, from scoping through deployment.
• Design and architect LLM training, fine-tuning, and evaluation pipelines tailored to client use cases.
• Translate client requirements into concrete technical architectures and delivery plans.
• Lead discovery sessions with client engineering and data science teams.
• Own model performance, evaluation methodology, and deployment recommendations for client-facing solutions.
• Collaborate with sales and delivery teams to ensure architectural decisions are sound and commercially deliverable.

Skills

LLM training

Transformer architectures

PyTorch/Hugging Face

Stakeholder communication

Job description

Job Title: Customer-Facing AI Data Scientist (LLM Architecture)

Role:

A leading UK Managed Service Provider (MSP), expanding its capability into applied AI and LLM-based solutions for enterprise clients, is hiring a Customer-Facing AI Data Scientist to sit at the intersection of deep technical LLM expertise and client delivery. This is not a back-office research role: you'll be in the room with client engineering leads and technical stakeholders, architecting solutions, defending design decisions, and translating training-level model knowledge into commercially viable outcomes for existing and prospective managed services clients. It suits someone who wants their technical depth to be visible and valued in front of the people who actually make buying and build decisions.

Genuine architect-level ownership: you're shaping how LLMs are fine-tuned, evaluated, and deployed for named enterprise accounts, not just executing tickets from a backlog.

Ground-floor opportunity to build out a new AI practice within an established, commercially stable MSP, rather than fighting for airtime inside a crowded AI-native startup.

A tight, technically elite peer group where "customer facing" doesn't mean diluting the science, it means being trusted to represent it.

Responsibilities:

• Act as the primary technical point of contact for enterprise clients on LLM-based engagements, from scoping through to deployment.
• Design and architect LLM training, fine-tuning, and evaluation pipelines tailored to specific client use cases and data environments.
• Translate ambiguous client requirements into concrete technical architectures and delivery plans.
• Lead deep technical discovery sessions and whiteboard sessions with client engineering and data science teams.
• Own model performance, evaluation methodology, and responsible deployment recommendations for client-facing solutions.
• Partner with sales, account management, and delivery teams to ensure architectural decisions are technically sound and commercially deliverable within a managed services model.

Skills/Must have:

• Experience: 5+ years in applied data science or ML engineering, with direct exposure to LLM training or fine-tuning at a technical (not just API-consumption) level.
• Core Tech/Domain: Hands-on experience with transformer architectures, LLM pre-training or fine-tuning workflows, and modern ML frameworks (PyTorch, Hugging Face, or equivalent).
• Methodology/Protocols: Strong grounding in evaluation methodology, RAG and fine-tuning trade-offs, and production ML deployment patterns.
• Soft Skills: Confident and credible in front of senior technical stakeholders, comfortable with pushback in live technical discussions, able to hold a boardroom and a terminal equally well.

Nice to haves:

• Experience with distributed training infrastructure (multi-GPU, multi-node).
• Prior pre-sales, solutions architecture, or consulting background.
• Familiarity with model safety, alignment, or responsible AI evaluation frameworks.
• Experience working within an MSP, systems integrator, or managed services delivery model.