A hands-on workshop for engineers building VLM-powered OCR that works on real-world financial documents.

Most document pipelines fail quietly. They work on a handful of PDFs, PPTs, XLS — then break the moment you scale across complex financial data with tables, charts, complex layouts, and more. Edge cases no amount of templating will ever account for.

FinParse NYC is a hands-on workshop hosted by Jerry Liu, CEO of LlamaIndex, where engineers take real, intentionally difficult financial documents and build a pipeline that helps them transform it into structured, usable data. We explore OSS alternatives as well as the value of using a commercial document ingestion API like LlamaParse that provides industry-leading accuracy, cost-efficiency and scale, as well as proprietary confidence scores and citations for HITL review.

This isn't a demo. You'll build a working pipeline on documents that are known to break things.

Agenda:

3:00-3:30 - Jerry Liu, CEO of LlamaIndex, presents on the future of agentic document OCR and how AI agents have transformed a trillion dollar legacy industry
3:30 - 5:00 - Hands on builders workshop
5:00 - 6:00 - Lessons from the field: How a Top 5 private equity firm built a risk decision making agent
6:00 PM - Dinner and Networking

What You'll Walk Away With

A working parsing + extraction pipeline you can extend
Practical schema design patterns for real-world financial documents
A clear understanding of why most document AI systems fail in production and what to do about it.

Location in midtown to be disclosed to all registrants shortly.

Turn Complex Financial Docs into LLM-ready Data with Jerry Liu and LlamaIndex