Back to Projects
Qira
Professional Project
Overview
Your AI Repair Assistant. Repairs in record time!
Qira is an AI-powered repair assistant developed by AHEAD GmbH which provides a web-based platform for workshops to quickly and accurately diagnose and repair vehicles. We leverage the vast databases of OEM repair manuals available under the right to repair legislation to provide step-by-step repair instructions, diagnostic tools, and parts identification. Qira aims to streamline the repair process, reduce downtime, and improve the overall efficiency of workshops by harnessing the power of AI to assist technicians in their daily tasks.
My Contribution
- Sole owner of the AI-powered repair assistant chatbot — responsible for architecture, development, and delivery end-to-end
- Developed across a NestJS (TypeScript) backend and Angular frontend on a live production system
- Built API integrations connecting 3 platforms (Ruby on Rails, NestJS, HubSpot), syncing 30,000+ records
- Set up PostgreSQL event tracking in Docker, migrating data from a legacy Ruby application across 3 environments
- Built a Python data scraping pipeline, indexing content into Elasticsearch to power ML document retrieval
- Integrated Azure Blob Storage for image persistence across the AI pipeline
- Engineered LLM prompt architecture — query refinement, document selection, and context injection for accurate AI responses
- Build agent-first with Claude Code: Opus/Fable in plan mode for architecture decisions, Sonnet for implementation, Git worktrees for parallel tickets, and a CLAUDE.md defining conventions every agent instance follows
- Led DevOps modernization: containerized environments in Docker, evaluating Docker Swarm for blue-green deployments and Ansible for infrastructure automation
- Self-taught systems administration managing Google Workspace, MDM tools, and company access systems
- Researched and configured Tableau dashboards for product analytics