AI automation spanning coding, PR review, and log analysis
Automation across writing, testing, reviewing, and monitoring code
Developers retain full control over approvals, merges, and releases
myRiva
San Mateo, CA
Travel Technology
Amazon Bedrock, AWS Lambda, Amazon S3, Bitbucket, Jenkins, Grafana, Graylog, New Relic, Jira, Confluence
MyRiva is a travel web application platform that enables customers to book travel reservations, including a white-label version for enterprise clients. The engineering team maintains a large, evolving codebase across frontend, backend, and middleware layers. Avahi partnered with MyRiva to design and deliver an AI-powered development automation capability that augments the engineering team across three core workflows: intelligent coding assistance for bug resolution and test generation, automated pull request review with vulnerability detection, and observability-driven log analysis for proactive issue identification. All AI-generated outputs operate in a human-in-the-loop mode, ensuring developers retain full control over approvals, merges, and releases.
MyRiva is a San Mateo, California-based travel technology company operating a web application platform for travel reservations. The platform serves both direct consumers and enterprise clients through a white-label offering. MyRiva’s engineering team manages a multi-layer application stack built on Java, JavaScript, and C#, with CI/CD workflows running through Bitbucket and Jenkins, and production monitoring across Grafana, Graylog, and New Relic.
MyRiva’s engineering team faced three compounding operational bottlenecks. Manual QA processes, bug fixing, and support ticket handling consumed significant developer time, with no automated path to generate tests from code changes or analyze support tickets at scale. Pull request reviews relied entirely on human reviewers, with no systematic detection of vulnerabilities, anti-patterns, or risky changes before code merged. And the observability stack, spanning three platforms, generated raw data that required manual interpretation to identify recurring errors, performance anomalies, and scaling risks.
These bottlenecks created a cycle of reactive engineering: developers spent time context-switching between bug fixes, test maintenance, and log analysis rather than building features. The objective was to use AI to augment developers across all three workflows, enabling faster resolution cycles while improving overall code quality and system reliability.
MyRiva’s existing infrastructure runs on AWS, making Amazon Bedrock the natural foundation for AI-powered development tooling. Bedrock provided access to large language models through a managed API without requiring MyRiva to host or fine-tune models independently, while AWS Lambda enabled serverless execution of AI-driven functions for test generation, code review, and log analysis. The entire solution deployed within MyRiva’s own AWS account, ensuring the company retained full control over its code, logs, and AI-generated outputs.
Avahi brought the combination of AWS technical depth and applied generative AI expertise needed to tackle a problem that spanned development tooling, CI/CD integration, and observability analytics. As an AWS Premier Tier Partner, Avahi had hands-on experience with Amazon Bedrock, AWS Lambda, and the integration patterns required to embed AI capabilities into existing engineering workflows without disrupting established processes. MyRiva needed a partner that could both architect the system and deliver working implementations across three distinct use cases within a compressed timeline, and Avahi’s track record in similar AI-assisted development engagements made them the right fit.
Avahi designed and delivered three interconnected AI-powered development automation capabilities, each integrated into MyRiva’s existing toolchain.
Use Case 1 — AI Coding Assistant: Avahi built an AI-powered coding assistant that analyzes support tickets, maps related code paths, proposes bug fixes with refactoring suggestions, and generates unit and regression tests to validate fixes. The assistant integrates with MyRiva’s IntelliJ-based development environment and operates through both an IDE-embedded copilot interface and a task-oriented coding agent for complex, multi-step code changes. Test generation is code-aware and change-driven, producing tests based on code structure, comments, annotations, and recent diffs.
Use Case 2 — PR Agentic Review: Avahi implemented an automated pull request review agent that triggers via Jenkins webhook when a PR is created in Bitbucket. The agent performs automated analysis to detect potential vulnerabilities, flag risky changes and anti-patterns, and suggest remediations aligned with best practices. Custom MCP integrations were built for Confluence and Bitbucket with scoped API tokens, enabling the agent to pull contextual documentation and repository information into its review process. All recommendations remain advisory, preserving the human review workflow.
Use Case 3 — Observability Log Analysis: Avahi built an observability analysis capability that ingests production logs and applies Named Entity Recognition (NER) combined with LLM-based classification to dynamically detect error patterns in booking logs. Unlike the existing approach — which relied on hardcoded conditional logic with no pattern matching — the AI-driven solution identifies recurring errors and anomalies, suggests potential root causes, and recommends areas for test coverage and refactoring. The system integrates with Jira for automated ticket creation when critical patterns are detected.
To address a platform constraint where certain Bedrock models were not accessible through MyRiva’s AWS account configuration, Avahi provided an alternative deployment path using a CLI-based agent framework, with documentation covering both options to ensure the customer could operate independently regardless of the model access resolution.
The engagement delivered a complete AI-assisted development automation layer that MyRiva’s engineering team can operate and extend independently. For the first time, the team has automated tooling that spans the full development lifecycle — from writing and testing code, to reviewing pull requests, to analyzing production logs — all integrated into their existing Bitbucket, Jenkins, and observability stack.
The customer characterized the engagement as “very well organized” and noted the solution provides “a cornerstone, a foundation for us to keep building.” The delivered capabilities position MyRiva to reduce manual QA effort, accelerate bug resolution cycles, and shift from reactive log monitoring to proactive, AI-driven issue detection. Follow-on work is under discussion to expand the solution across additional services and workflows.
myRiva
San Mateo, CA
Travel Technology
Amazon Bedrock, AWS Lambda, Amazon S3, Bitbucket, Jenkins, Grafana, Graylog, New Relic, Jira, Confluence
MyRiva is a travel web application platform that enables customers to book travel reservations, including a white-label version for enterprise clients. The engineering team maintains a large, evolving codebase across frontend, backend, and middleware layers. Avahi partnered with MyRiva to design and deliver an AI-powered development automation capability that augments the engineering team across three core workflows: intelligent coding assistance for bug resolution and test generation, automated pull request review with vulnerability detection, and observability-driven log analysis for proactive issue identification. All AI-generated outputs operate in a human-in-the-loop mode, ensuring developers retain full control over approvals, merges, and releases.
MyRiva is a San Mateo, California-based travel technology company operating a web application platform for travel reservations. The platform serves both direct consumers and enterprise clients through a white-label offering. MyRiva’s engineering team manages a multi-layer application stack built on Java, JavaScript, and C#, with CI/CD workflows running through Bitbucket and Jenkins, and production monitoring across Grafana, Graylog, and New Relic.
MyRiva’s engineering team faced three compounding operational bottlenecks. Manual QA processes, bug fixing, and support ticket handling consumed significant developer time, with no automated path to generate tests from code changes or analyze support tickets at scale. Pull request reviews relied entirely on human reviewers, with no systematic detection of vulnerabilities, anti-patterns, or risky changes before code merged. And the observability stack, spanning three platforms, generated raw data that required manual interpretation to identify recurring errors, performance anomalies, and scaling risks.
These bottlenecks created a cycle of reactive engineering: developers spent time context-switching between bug fixes, test maintenance, and log analysis rather than building features. The objective was to use AI to augment developers across all three workflows, enabling faster resolution cycles while improving overall code quality and system reliability.
MyRiva’s existing infrastructure runs on AWS, making Amazon Bedrock the natural foundation for AI-powered development tooling. Bedrock provided access to large language models through a managed API without requiring MyRiva to host or fine-tune models independently, while AWS Lambda enabled serverless execution of AI-driven functions for test generation, code review, and log analysis. The entire solution deployed within MyRiva’s own AWS account, ensuring the company retained full control over its code, logs, and AI-generated outputs.
Avahi brought the combination of AWS technical depth and applied generative AI expertise needed to tackle a problem that spanned development tooling, CI/CD integration, and observability analytics. As an AWS Premier Tier Partner, Avahi had hands-on experience with Amazon Bedrock, AWS Lambda, and the integration patterns required to embed AI capabilities into existing engineering workflows without disrupting established processes. MyRiva needed a partner that could both architect the system and deliver working implementations across three distinct use cases within a compressed timeline, and Avahi’s track record in similar AI-assisted development engagements made them the right fit.
Avahi designed and delivered three interconnected AI-powered development automation capabilities, each integrated into MyRiva’s existing toolchain.
Use Case 1 — AI Coding Assistant: Avahi built an AI-powered coding assistant that analyzes support tickets, maps related code paths, proposes bug fixes with refactoring suggestions, and generates unit and regression tests to validate fixes. The assistant integrates with MyRiva’s IntelliJ-based development environment and operates through both an IDE-embedded copilot interface and a task-oriented coding agent for complex, multi-step code changes. Test generation is code-aware and change-driven, producing tests based on code structure, comments, annotations, and recent diffs.
Use Case 2 — PR Agentic Review: Avahi implemented an automated pull request review agent that triggers via Jenkins webhook when a PR is created in Bitbucket. The agent performs automated analysis to detect potential vulnerabilities, flag risky changes and anti-patterns, and suggest remediations aligned with best practices. Custom MCP integrations were built for Confluence and Bitbucket with scoped API tokens, enabling the agent to pull contextual documentation and repository information into its review process. All recommendations remain advisory, preserving the human review workflow.
Use Case 3 — Observability Log Analysis: Avahi built an observability analysis capability that ingests production logs and applies Named Entity Recognition (NER) combined with LLM-based classification to dynamically detect error patterns in booking logs. Unlike the existing approach — which relied on hardcoded conditional logic with no pattern matching — the AI-driven solution identifies recurring errors and anomalies, suggests potential root causes, and recommends areas for test coverage and refactoring. The system integrates with Jira for automated ticket creation when critical patterns are detected.
To address a platform constraint where certain Bedrock models were not accessible through MyRiva’s AWS account configuration, Avahi provided an alternative deployment path using a CLI-based agent framework, with documentation covering both options to ensure the customer could operate independently regardless of the model access resolution.
The engagement delivered a complete AI-assisted development automation layer that MyRiva’s engineering team can operate and extend independently. For the first time, the team has automated tooling that spans the full development lifecycle — from writing and testing code, to reviewing pull requests, to analyzing production logs — all integrated into their existing Bitbucket, Jenkins, and observability stack.
The customer characterized the engagement as “very well organized” and noted the solution provides “a cornerstone, a foundation for us to keep building.” The delivered capabilities position MyRiva to reduce manual QA effort, accelerate bug resolution cycles, and shift from reactive log monitoring to proactive, AI-driven issue detection. Follow-on work is under discussion to expand the solution across additional services and workflows.
Let’s explore your high-impact AI opportunities together in a complimentary session
