Large Language Models in Finance

Financial AI has to do more than produce fluent answers: it must handle numbers carefully, work with changing evidence, and remain subject to meaningful controls. Large Language Models in Finance takes that practical challenge as its starting point, connecting the mechanics of LLMs with the demands of building and governing financial systems.

Miquel Noguer i Alonso moves from transformer foundations and prompting into fine-tuning, retrieval-augmented generation (RAG), and agent-based workflows. Along the way, the book keeps attention on evaluation, reliability, auditability, and human oversight—not just what a model can do, but how its use can be tested and controlled.

From model foundations to financial applications

The early chapters introduce language-model architectures, tokenization, embeddings, training, and financial reasoning. That technical groundwork leads into hands-on topics including fine-tuning methods such as LoRA and QLoRA, RAG pipelines for financial information, and agents that use tools and retrieval within structured workflows.

Build systems with evidence and controls

Rather than treating deployment as an afterthought, the book considers evaluation, security, monitoring, model risk, and governance alongside implementation. Its coverage includes financial research, trading support, banking operations, fraud, risk, and compliance, with case studies that bring system-design choices into view.

• Explore transformer architecture, prompting, and financial model design.
• Study fine-tuning, RAG, and financial agent workflows.
• Consider evaluation, numerical reliability, auditability, and oversight.
• Connect technical methods to financial use cases and Python-based implementation.

For practitioners working at the intersection of AI and finance

This title is relevant to data scientists, quantitative analysts, AI and fintech developers, banking and compliance professionals, researchers, and graduate students seeking to apply LLMs in financial contexts. The publisher indicates that readers should have a foundation in Python programming, machine learning, and financial markets.

For readers designing financial AI, the central thread is clear: useful systems depend on more than capable models. They also need sound data practices, careful evaluation, and controls suited to the work they perform.