Data Scientist - Generative AI / Senior Consultant Specialist

• Design and deliver end-to-end Machine Learning solutions, from problem framing and data exploration through to production deployment and monitoring.
• Build LLM-powered applications using frameworks such as LangChain and/or LlamaIndex, aligned to real business use cases.
• Develop and optimise prompt engineering approaches to improve response quality, consistency, and reliability.
• Implement Retrieval-Augmented Generation (RAG) pipelines using vector databases such as FAISS, Pinecone, Chroma, or Weaviate.
• Fine-tune and adapt pre-trained LLMs (e.g., GPT, LLaMA, Mistral, Claude, Gemini) on domain-specific datasets where required.
• Create robust NLP pipelines for tasks such as NER, text classification, sentiment analysis, information extraction, summarisation, and Q&A.
• Leverage the Hugging Face Transformers ecosystem to evaluate, customise, and productionise NLP/LLM models.
• Build and expose models and AI services via REST APIs using FastAPI and/or Flask, following secure and maintainable engineering practices.
• Collaborate with MLOps/DevOps to containerise and deploy solutions (e.g., Docker, CI/CD) and support production operations.
• Communicate insights and solution outcomes clearly to technical and non-technical stakeholders, using visualisations and crisp storytelling

To be successful, you will:

• Strong hands-on expertise in Python and the data science ecosystem (pandas, NumPy, scikit-learn).
• Proven experience delivering NLP solutions using libraries such as Hugging Face Transformers, spaCy, and/or NLTK.
• Practical experience implementing LLM use cases (prompting, RAG, evaluation), ideally with LangChain/LlamaIndex and LLM APIs.
• Solid understanding of ML fundamentals: feature engineering, model selection, hyperparameter tuning, cross-validation, and performance evaluation.
• Hands-on deep learning experience with PyTorch or TensorFlow.
• Experience working with structured and unstructured data, including strong SQL capability and familiarity with vector databases.
• Ability to validate and control LLM outputs using structured techniques (e.g., Pydantic, output parsers) and reduce hallucination risk.
  • Strong stakeholder management and communication skills—able to explain complex concepts simply and drive delivery in a collaborative team setup.