Nebula Analytics
Nebula needed an analytics product that felt instant even when slicing billions of events. We built a dashboard that streams results as you type, with WebGL-accelerated charts that stay at 60fps.
- Client
- Nebula (Series B SaaS)
- Role
- Lead Frontend Engineer
- Timeline
- 8 weeks
- Year
- 2025
The challenge
The legacy dashboard took 8–12 seconds to render a single report, and customers were churning over it. Queries were heavy, charts re-rendered the whole DOM, and there was no perceived progress.
The approach
We moved aggregation to a columnar store at the edge, streamed partial results over a binary protocol, and rendered charts on the GPU. A virtualization layer kept the DOM tiny while a Suspense-driven UI showed instant skeletons.
The outcome
Median report time dropped from ~9s to under 800ms, engagement on the analytics tab tripled, and Nebula closed three enterprise deals citing speed as the deciding factor.
What I did
- Architected the streaming query → render pipeline
- Built the WebGL charting layer and virtualization
- Set the performance budget and CI perf gates
- Mentored two engineers through the rebuild
Tech stack
Have a bold idea? Let's build it.
I take on a handful of projects each quarter. Tell me what you're dreaming up and let's make it real.