Machine Learning SDLC Anomaly Detection with Adaptive Actions

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional SDLC methods face challenges in achieving comprehensive end-to-end traceability and risk management in distributed software environments, leading to software delivery risks, quality impacts, and increased likelihood of production incidents due to inconsistent data collection and manual processes.

Innovation Solution

A machine learning-powered system integrates data from various SDLC tools, provides a single pane of glass view, and generates actionable recommendations using a mixed pipeline of machine learning models and adaptive models to identify bottlenecks, risks, and outliers, with the capability to execute recommended actions or a kill switch when necessary.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional manual processes are used for SDLC data collection, then ease of operation is maintained, but measurement precision and reliability of traceability data deteriorate

Engineering Contradiction:

Improvetraceability data accuracyVSAvoidmanual process complexity

Core Design Contradiction:

Measurement precisionVSEase of operation

Solution Approach 1:

The patent replaces manual mechanical data collection processes with automated machine learning models that ingest data from multiple SDLC tools (JIRA, GitHub, Jenkins, etc.). The system automatically traces user stories through the SDLC pipeline, eliminating manual tracking while improving data accuracy and consistency.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service data collection by automatically connecting to and ingesting data from various SDLC tools without requiring manual intervention. The machine learning models autonomously process, analyze, and trace data through the software development lifecycle, reducing operational burden while enhancing measurement precision.

Inventive Principle:

Principle #25Self-service

2Reliability

If comprehensive end-to-end traceability is implemented across all SDLC tools, then reliability of risk management improves, but device complexity increases

Engineering Contradiction:

Improverisk management reliabilityVSAvoidsystem integration complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the complex traceability system into distinct machine learning models, each responsible for specific SDLC phases or tools. This modular architecture handles integration complexity independently while maintaining comprehensive end-to-end traceability across requirements, design, implementation, testing, and deployment stages.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system introduces an intermediary machine learning layer that sits between multiple SDLC tools and the traceability analysis function. This intermediary automatically ingests, standardizes, and processes data from diverse sources (JIRA, GitHub, Jenkins, etc.), simplifying integration while ensuring reliable cross-tool traceability.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Productivity

If real-time anomaly detection is implemented using machine learning models, then productivity through automated insights improves, but use of energy and computational resources increases

Engineering Contradiction:

Improveautomated insight generationVSAvoidcomputational resource consumption

Core Design Contradiction:

ProductivityVSUse of energy by moving object

Solution Approach 1:

The system applies partial anomaly detection by focusing machine learning models on specific high-risk SDLC phases or critical data points rather than analyzing all data uniformly. This selective approach generates actionable insights where most needed while reducing overall computational resource consumption.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The machine learning models implement feedback mechanisms that learn from detected anomalies and adjust their analysis focus dynamically. The system refines its computational efforts based on historical patterns and risk assessments, maintaining high productivity in anomaly detection while optimizing energy usage by concentrating resources on high-priority areas.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS12423656B2Machine learning-based anomaly detection and action generation system for software development lifecycle management

Publication Date: 2025.09.23 CITIBANK N A

AI summary

Systems and methods for enhanced software development lifecycle management using a pipeline of machine learning and adaptive models for anomaly detection and action generation. In some aspects, the system may generate a system data stream for an SDLC management platform that stitches together source data from multiple sources, generates a query to a first machine learning model for a user story identifier to generate trend information and anomalies, based on output from a second machine learning model including dynamic thresholds for the anomalies, determining at least one anomaly that satisfies a corresponding dynamic threshold, processing using an adaptive model the at least one anomaly to generate recommended actions to address the at least one anomaly, and generating for inclusion in the graphical user interface the one or more recommended actions.