BlogFrom the team
From the team
building Veldora.
Engineering deep dives, product thinking, and lessons learned shipping a multi-tenant AI platform.
Featured · Engineering
How we cut chatbot latency by 41% with hybrid retrieval
A pragmatic look at where we spent the milliseconds, why pure vector search wasn’t enough, and what we changed in production.
By Shubham · May 14, 2026 · 9 min read
RAG
Product
Why Veldora answers only from your site
The product philosophy behind grounded-only chatbots — and how we enforce it technically.
By Shubham · May 02, 2026 · 5 min read
Engineering
A practical guide to chunking websites
Real-world heuristics for breaking marketing, docs, and product pages into retrieval-ready chunks.
By Team · Apr 22, 2026 · 8 min read
Architecture
Why we run scraping on EC2, not serverless
When long-running, IO-heavy workers beat Lambda economics. Numbers included.
By Team · Apr 10, 2026 · 6 min read
Security
Multi-tenant isolation, end to end
How the Veldora platform keeps every customer’s data, vectors, and conversations strictly separate.
By Team · Apr 01, 2026 · 7 min read