jenis donda@jenis_donda

CyberGuardian autonomously collects and delivers the latest information from diverse news sources. We use the Trulens for evaluating and tracking our model responses. It provides programmatic methods for generating different metrics(answer relevance, groundedness, context relevance) on an application run. We have built a scheduler that runs on specific intervals. It fetches real-time news, articles, and blogs regarding to cyber security domain. We store this real-time data in the Mongo DB as vectors. When the user hits a query, our app fetches relevant vectors related to questions and passes them to the LLM model. Then trulens run after that and provides different metrics including costing, no token use with his request, answer relevance, groundedness, and context relevance.

As the software development cycle speeds up and the codebase grows, code review is vital in enhancing code quality, identifying bugs, and maintaining code standards. To solve that, CodeSensei is a comprehensive Best Practice Analyzer designed to enhance code quality. Users submit repo link to their code and corresponding best practice documents, and CodeSensei generates a detailed analysis report. The report highlights areas needing improvement with visual indicators and actionable suggestions based on the best practices. It includes color-coded highlights and annotated comments to help users understand how to align their code with coding standards and improve overall quality.