A conversational AI assistant integrated into Smarteeva’s Docusaurus documentation ecosystem. It helps users instantly understand complex product concepts, find answers, and navigate technical documentation through natural-language queries. Powering this experience is a hybrid RAG architecture, combining static embeddings, dynamic updates, and reference-driven responses for trust and accuracy.

A full-stack automation platform that enables technical and non-technical users to generate stable, production-grade Playwright tests using either natural language or recorded browser scripts. The system enhances raw scripts with AI, normalizes unstable Salesforce selectors, runs tests on GitLab CI, and returns full video + HTML reports in a clean UI. Built with Node, Playwright, GitLab Pipelines, Docker, and internal design system components.

A unified, AI-powered internal tool that generates, structures, and exports release notes automatically from Jira stories and multi-package sprint data. Designed for teams struggling with inconsistent manual notes and decentralized workflows across different Salesforce-based product packages.

A React Native mobile app that enables field teams, quality specialists, and support staff to submit, review, and track MDR-related complaints instantly from anywhere. Designed to reduce reporting delays, simplify data entry, and modernize Smarteeva’s complaint-management workflow for on-site use.

A complete end-to-end redesign of Smarteeva’s internal Admin Portal — including dashboards, data tables, notifications, billing, user management, MDR search, and foundational UI patterns. The new design system streamlined daily operations, reduced navigation steps, and established a scalable enterprise interface for future AI-driven tools.

A complete transformation of Smarteeva’s legacy decision-tree tool into a modern, visual rule-building environment with instant preview, guardrails, and intuitive branching. The redesign reduced configuration errors, accelerated admin workflows, and became the core logic engine powering multiple regulatory modules.

A unified platform for testing, comparing, and validating LLM outputs across multiple models and environments. Designed to help engineering, QA, and R&D teams benchmark model performance, diagnose failures faster, and centralize all prompt testing workflows into one predictive, visual system.