ELC
Sign in or Join the community to continue

Mastering Unstructured Data at Scale

Posted Oct 29, 2026 | Views 0
Share

Speaker

user's Avatar
Ruchita Chandhok
Senior Director of Engineering @ Carta

From global capital markets to enterprise AI, Ruchita Chandhok has spent her career building the systems behind some of the world’s most demanding financial platforms. As a Senior Director of Engineering at Carta, she leads global engineering organizations at the intersection of financial infrastructure and artificial intelligence.

Across public and private markets, the platforms she has helped build have managed more than half a trillion dollars in financial assets. Today, she is focused on applying generative AI to one of enterprise software’s hardest challenges transforming complex, unstructured financial data into trusted systems that power critical business decisions.

Ruchita’s perspective is rooted in a simple belief: the next generation of software will be defined not by AI models alone, but by the engineering systems built around them. She speaks about architecting production-grade AI, maximizing business value through intelligent token utilization, and designing scalable platforms where every AI interaction delivers measurable outcomes with the reliability, explainability, and trust demanded by enterprise financial systems.

+ Read More

SUMMARY

Onboarding a new Carta customer meant eight manual steps: emails, spreadsheets, re-uploads, and an implementation manager stitching a cap table together by hand from dozens of legal PDFs. It is now one drop-in flow, at 97% precision and 94% recall on extraction from unstructured documents. A firm arriving with 8,000 documents across hundreds of portfolio companies now onboards in weeks instead of months.

Ruchita Chandhok uses that build to work through four decisions any team putting an LLM in front of real data will face:

Where the contract between platform and product lives. Typed schemas, versioned like APIs, not prompts. It is what lets the two teams stop blocking each other. What a reviewer needs to trust a machine-extracted value. Provenance on every field, down to the page and sentence. Verification in seconds beats confidence scores. When to build evals. Before the second feature, not after the first incident. You cannot reconstruct a regression from logs. What the model should never be allowed to do. Arithmetic, dedup, and date math belong in tested code. Reserve the model for judgment.

+ Read More

CONTENT & TRANSCRIPT


+ Read More
Sign in or Join the community
Exclusive to engineering leaders
ELC
Create an account
Exclusive to engineering leaders
Comments (0)
Popular
avatar


Watch More

Building a Culture of Execution at Scale by Scott Shumaker
Posted Nov 26, 2020 | Views 1K
# ELC Annual 2020
# Scaling
# Collaboration
# Leadership
# Execution
Technology Decisions At Scale Using ADRs & Tech Radars by Yonas Beshawred
Posted Nov 02, 2020 | Views 799
# Technology
# ELC Annual 2020
# DevOps
# Leadership
From Awkward to Agonizing - Mastering Difficult Conversations
Posted Apr 01, 2019 | Views 840
# Career
# Collaboration
# Communication
# Culture
# Decision Making
# Influence
# Leadership
# Management
# ELC Summit 2019