Scribd
Upload
72 views

Flipkart

The document proposes using large language models to develop a system for compliance monitoring and enforcement through log analysis. The system would analyze logs, configurations, access controls, and privileges to check for compliance with security policies and standards. By leveraging models like ChatGPT, the system would automate identifying non-compliant activities and generate insights for remediation. Key attributes of the proposed system include flexible handling of data formats and standards, actionable insights, scalability, adaptability, and high accuracy in identifying compliance breaches. The workflow would involve providing compliance documents and logs to the system, which would then detect breaches and provide insights to address issues. An open source LLM with a generic automation stack is suggested technically.

Uploaded by

Rom On
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
72 views

Flipkart

The document proposes using large language models to develop a system for compliance monitoring and enforcement through log analysis. The system would analyze logs, configurations, access controls, and privileges to check for compliance with security policies and standards. By leveraging models like ChatGPT, the system would automate identifying non-compliant activities and generate insights for remediation. Key attributes of the proposed system include flexible handling of data formats and standards, actionable insights, scalability, adaptability, and high accuracy in identifying compliance breaches. The workflow would involve providing compliance documents and logs to the system, which would then detect breaches and provide insights to address issues. An open source LLM with a generic automation stack is suggested technically.

Uploaded by

Rom On
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Compliance Monitoring and Enforcement through Log Analysis using Large

Language Models ( Infosec Engineering)

Problem Statement :
As Flipkart deals with increasing volumes of data and complex systems, ensuring compliance
with security policies, standards and baselines has become a critical challenge. To address this
issue, we propose a project focused on developing a system that leverages large language
models for compliance monitoring and enforcement through log analysis from relevant sources.

The objective is to build a solution that can effectively analyze logs, system configurations,
access controls, and user privileges to check for compliance with security policies and
standards. By utilizing the power of large language models (LLMs) like ChatGPT or its open
source alternatives, we aim to automate the process of identifying non-compliant activities and
generating actionable insights for remediation.

—-------------------------------

Here are the following attributes that we have to keep in regard while developing this.

Solution Attributes of Compliance Monitoring and Enforcement :

● Rule Definition: The ability to distill the rules and use the standards provided to infer the
possible relationships between rules and parameters in the logs/policies.
● Flexibility: The ability of the system to handle various log formats (CSV, text, PDF), rule
sets, and compliance standards. It should be adaptable to different environments and
configurations.
● Actionable Insights: The relevance and usefulness of the actionable insights provided
by the system for remediation purposes
● System Performance: The efficiency and scalability of the system in handling large
volumes of logs and text data.
● Adaptability: The system's capability to learn from new rules, update compliance
standards, and incorporate feedback for continuous improvement without restarting
. ● High Accuracy/Precision/Recall: The ability to identify compliance breaches along with
low false positives and and very low false negatives.
● UI/UX (Optional): a way to interact with the system, configure rules, upload logs and
text files, view compliance reports, and access actionable insights.

Application workflow is that we provide some compliance documents as ruleset along with
logs/system policies etc with known failures. Then we should be provided with the compliance
breaches along with citations to the specific line/log file/userid. The system should also provide
actionable insights i.e. what can be done to fix the issues.
Technical Requirements - An open source LLM (APIs not permitted because they send data to
third party servers) + generic tech stack for automation and parsing.

You might also like