Principal AI Consultant / Data Scientist
Principal AI Consultant / Data Scientist
The AI Practice helps customers and prospects access senior AI and Data Science expertise without having to recruit a permanent AI team of their own. The Principal AI Consultant is a billable specialist within the practice: a hands-on, customer-facing NLP/LLM practitioner who can join client conversations, provide technical discovery and specialist expertise, and build the proof of concept or proof of value that moves the opportunity forward.
Customers may receive defined blocks of specialist work, typically including use case analysis, benchmarking, RAG or fine-tuning assessments, agent design, or a working prototype, each with a documented outcome. The Head of AI Practice works alongside the Principal AI Consultant on broader discovery, business framing and solution scoping, while the Principal AI Consultant provides the technical discovery, analysis and specialist AI/Data Science expertise. Where additional software, data or platform engineering capability is required, the appropriate ITPS teams or specialist partners are brought into the engagement.
The practice can develop proofs of concept and lightweight AI applications using modern AI platforms and services, while larger production builds are delivered with software engineering, data engineering, platform engineering or specialist partner support.
What You'll Deliver
- Use case analysis – working with client stakeholders and the Head of AI Practice to identify and scope viable AI/NLP use cases.
- Model benchmarking – evaluating candidate models and approaches against client data and success criteria.
- RAG decisions – assessing retrieval-augmented generation architectures and recommending the right approach.
- Fine-tuning assessment – determining when fine-tuning a pre-trained model is justified over prompting or RAG, and scoping the work.
- Agent design – designing agentic workflows and orchestration for client use cases.
- Prototype development – building working prototypes and proofs of concept that demonstrate value quickly.
- Vendor evaluation – assessing third-party models, platforms and tooling on a client’s behalf.
- Support to internal software teams – advising client engineering teams on taking AI components into production.
What Good Looks Like - Experience and Skills
Technical depth
- Deep, hands-on experience with modern NLP, LLM and generative AI systems, including transformer architectures, embedding models, retrieval-augmented generation, model evaluation, fine-tuning and inference.
- Practical experience working with both open-source and commercial foundation models, with an understanding of the strengths, limitations and trade-offs between different approaches.
- Strong practical experience with modern AI/ML frameworks and tooling such as PyTorch, Hugging Face, TensorFlow/Keras or equivalent, with the judgement to select the right approach for the client problem rather than defaulting to the most familiar tool.
- Experience designing and building LLM applications, including prompting, structured outputs, embeddings, vector search, RAG and agent/tool-based workflows.
- Understanding of the considerations involved in moving AI applications beyond prototype, including evaluation, security, guardrails, performance, cost and operational requirements.
- Comfortable operating across the full lifecycle from unstructured data processing and modelling through to prototype and, where relevant, production handover.
Client-facing capability
- Track record leading teams to deliver AI solutions directly to enterprise clients, not just internal projects.
- Able to translate a business problem into a scoped, defensible piece of technical work and translate the results back into terms a non-technical sponsor will act on.
- Credible in the room for pre-sales and early-stage scoping conversations, as well as delivery.
Way of working
- Comfortable with billable, outcome-defined engagements rather than open-ended staff augmentation.
- Pragmatic about when to build in-house versus bring in a partner or specialist team.
