LLM Researcher

Hybrid Working Required - 3 Days On, 2 days Remote Our client are an AI-Native FinTech start-up based in London. Brief:We're building an AI-native financial operations platform. As an ML/LLM Engineer, you'll design, fine-tune, and deploy models that automate complex financial workflows-payouts, reconciliations, compliance checks, and forecasting-while keeping humans in the loop for accuracy and trust. 

Responsibilities:

• Fine-tune and optimize foundation models (e.g., LLaMA, FinBERT) for financial text and transaction data.
• Build retrieval-augmented generation (RAG) pipelines with Hugging Face, Lang Chain, and vector databases.
• Implement human-in-the-loop systems to continuously validate, improve, and govern AI outputs.
• Deploy and monitor ML models in production with a focus on scalability, latency, and compliance
• .Collaborate with product and finance experts to translate real-world workflows into AI agents. 

Key Requirements:

• Strong experience in NLP/LLM training and fine-tuning (transformers, embeddings, LoRA/PEFT).Proficiency with PyTorch/TensorFlow, Hugging Face ecosystem, and MLOps tooling.
• Experience working with financial or structured transactional data.
• Knowledge of data privacy and regulatory requirements in fintech (SOC2, PCI-DSS, GDPR).
• Ability to thrive in a fast-moving startup environment. 

Desirables:

• Background in fintech, payments, or treasury systems.
• Experience with reinforcement learning with human feedback (RLHF).