LLMs & Agents in Finance

Read these first

1. 01Where LLMs Actually Help in Quant Workflows

   The honest answer is narrower than the hype, LLMs are strong at reading text at volume and weak at anything that requires actually knowing what a price will do.

2. 02Agentic Research Workflows

   An agentic workflow lets an LLM plan its own sequence of tool calls, search, extract, compute, instead of following one fixed prompt, trading predictability for the ability to handle open-ended questions.

3. 03Building a RAG Pipeline Over a Financial Corpus

   Retrieval-augmented generation lets a model answer from documents it can actually cite instead of from memory, the hard part is retrieval quality, not the generation step everyone focuses on.

4. 04Structured Extraction From Filings With LLMs

   Turning a 10-K or transcript into clean numeric fields is a search-and-verify problem, not a generation problem, treat the model as a fast, fallible reader and check its work.

Then the rest

1. 05Building an Evaluation Set for Financial LLM Tasks
2. 06Chunking Strategies for Long Filings
3. 07Compliance and Record-Keeping for LLM Use
4. 08Designing Tool Schemas for Quant Agents
5. 09Document Parsing Pipelines for Financial Corpora
6. 10Earnings-Call Q&A With LLMs
7. 11Guardrails and Action Limits for Agents
8. 12Hallucination Controls for Numeric Answers
9. 13Hybrid Retrieval: BM25 Plus Embeddings
10. 14Multi-Agent Research Pipelines
11. 15Point-in-Time Corpora for LLM Backtests
12. 16Reranking Retrieved Passages
13. 17Schema-Constrained JSON Output
14. 18Scoring Management Tone and Evasiveness
15. 19Table Extraction From Financial PDFs
16. 20Turning LLM Outputs Into Model Features