← All speakers
  • DataWig: Missing Value Imputation for Tables ↗

    Naidu co-authored the 2019 paper introducing a tool that combines deep-learning feature extractors and automated hyperparameter tuning to impute missing values in heterogeneous tables.

    jmlr.org
  • DataWig ↗

    Source code for the tool described in Naidu’s co-authored paper; the retained evidence does not establish his individual code contributions.

    github.com · Source ↗
  • Introducing Rutter Instant Link ↗

    Naidu’s 2022 announcement describes URL-based account authentication that can reduce lending application friction without a code integration.

    rutter.com
  • Introducing Rutter Write Accounting ↗

    Naidu’s 2022 announcement describes a unified API for writing standardized financial data across accounting platforms, including automated payment reconciliation.

    rutter.com

Bio, Work & Ideas

Prathik Naidu

Conference affiliation: Long Lake

Prathik Naidu is a founding partner at Long Lake, where he builds and deploys AI across the services economy. His work spans machine-learning research, customer-facing engineering, and products that automate business workflows.

Long Lake acquires and partners with services businesses, using its Nexus AI transformation platform to improve their operations. In 2026, it completed its acquisition of American Express Global Business Travel, with plans to apply AI to customer tools and support travel counselors.

From machine learning to business products

Naidu earned bachelor’s and master’s degrees in computer science at Stanford and built real-time graphics systems at NVIDIA. He was the first engineer at Skiff, later acquired by Notion. At Scale AI, he was a forward-deployed engineer leading customer-facing product and engineering development for Scale Document, which applied machine learning to document processing.

As a product lead at Rutter, Naidu worked on software connecting financial products to business data. He introduced Rutter Instant Link, which let a lender send a URL through which a business could connect its accounting, commerce, or payments accounts without the lender integrating code. The goal was to reduce friction in a loan application while giving the lender data for underwriting.

He also introduced Rutter Write Accounting, which used a unified API to send standardized data back to accounting systems. A payments company, for example, could record a payment and reconcile it with a bill without asking a finance team to re-enter the information.

Naidu co-authored the 2019 paper introducing DataWig, a tool for filling in missing values in tables. It combined deep-learning feature extractors with automated hyperparameter tuning to handle mixed data types, including unstructured text, with less manual feature engineering.

Read the topics behind these talks

1 conference talk

Key ideas

Scroll to read ↓

Prathik Naidu explains Long Lake’s case for acquiring established services businesses, deploying AI with their employees, and turning saved time into business results.

  • Enterprise deployment requires integration, adoption and a way to capture value; records alone do not reveal the judgment and exceptions embedded in everyday work.
    0:43 ↗
  • Acquiring an established service business can bring builders closer to experienced employees, live feedback, customer trust and shared business incentives.
    6:45 ↗
  • The proposed economic gain is operating leverage: serving more demand without labor costs rising as directly with revenue, while maintaining expert service.
    10:15 ↗
  • Time savings are an intermediate milestone. Management must put freed capacity to work before deployment can produce the intended P&L and service outcomes.
    16:02 ↗

References