Insights
The VextoraTech engineering blog
We write about what we build. AI, full-stack patterns, DevOps, and the occasional hard lesson.

Featured · AI & ML
Building a RAG Pipeline from Scratch with ChromaDB and LLaMA 3.2
How we built a local-first RAG system with citations, embeddings, and zero API cost.
Farjad · Jun 5, 2026 · 8 min read
Read article AI & ML
8 min readBuilding a RAG Pipeline from Scratch with ChromaDB and LLaMA 3.2
How we built a local-first RAG system with citations, embeddings, and zero API cost.
FarjadJun 5, 2026
Web Dev
6 min readWhy We Use the Repository Pattern in Every FastAPI Project
The architectural pattern that keeps our backends testable, swappable, and sane.
TaimoorMay 22, 2026
Web Dev
7 min readRBAC Done Right: 4 Roles, 16 Permissions, Zero Confusion
A pragmatic role-based access control schema you can ship on Monday.
FarjadMay 10, 2026
AI & ML
5 min readLocal AI vs. API: When to Use Ollama Instead of OpenAI
A cost, latency, and privacy comparison from real client projects.
MateenApr 28, 2026
DevOps
9 min readDocker Compose for Full-Stack Projects: Our Production Template
The compose file we copy into every project, annotated.
TaimoorApr 14, 2026
Design
6 min readDesigning for Developers: Building UI That Engineers Actually Use
Lessons from designing dashboards used by engineering teams.
FarjadApr 3, 2026
AI & ML
7 min readMermaid.js + AI: Generating Diagrams from Natural Language
How DiagramAI Studio turns one sentence into a system diagram.
FarjadMar 19, 2026
Web Dev
10 min readJWT Auth in FastAPI: Our Battle-Tested Implementation
Refresh tokens, rotation, and revocation — the production setup.
TaimoorMar 8, 2026






