About Nymbus
A Steady Lens on
Compute Performance
We build structured reference materials and planning tools so engineering teams can work with compute data without needing to second-guess its provenance.
Back to HomeOur Story
How Nymbus Came Together
Nymbus started in Petaling Jaya when a small group of infrastructure engineers kept running into the same problem: the compute performance figures they found online were either vendor-produced, poorly documented, or measured under conditions that bore little resemblance to real planning scenarios. There was no neutral, readable place to go when preparing an internal briefing.
The founders spent the better part of a year building a methodology for sourcing, recording, and presenting hardware figures in a way that would hold up to scrutiny inside a planning meeting. The result became the first version of what is now the Benchmark Reference Library.
Today Nymbus serves analysts and engineering teams across Malaysia who want a structured, interpretable information resource rather than a promotional datasheet. Every product we offer reflects the same underlying principle: clear framing makes better conversations possible.
Our Mission
Structured Data, Honest Framing
The Team
People Behind Nymbus
A small, focused team with backgrounds in infrastructure engineering, data analysis, and technical writing.
Suraya Norzaini
Co-Founder & Research Lead
Spent eight years in infrastructure planning at enterprise scale. Developed the source methodology that underpins every benchmark entry in the library.
Farid Kamarudin
Co-Founder & Product
Built tooling for technical teams at two Malaysian software firms before joining Nymbus. Leads product design for the Comparison Toolkit.
Lim Hui Wen
Head of Data & Methodology
Brings a background in applied statistics and hardware evaluation. Oversees quarterly refresh cycles and ensures figures meet documentation standards before publication.
How We Work
Quality Standards & Protocols
The practices that keep our data useful and trustworthy for teams who rely on it during real planning conversations.
Documented Source Methodology
Each benchmark entry is sourced according to a published methodology covering test environment, measurement approach, and applicable hardware scope.
Quarterly Review Cycle
All entries are reviewed every quarter. Figures that no longer reflect current conditions are revised or retired. Subscribers are notified of material changes.
Data Privacy Compliance
Subscriber data is handled in accordance with Malaysia's Personal Data Protection Act 2010. No usage data is shared with hardware vendors or third-party advertisers.
Clear Interpretation Guidance
Every chart and table includes a plain-language note explaining what the figure shows, what conditions apply, and where it should not be used as a sole basis for decisions.
No Vendor Influence Policy
Nymbus operates without commercial arrangements with any hardware or platform vendor. Content selection and presentation are made independently by our research team.
Responsive Client Support
Subscribers and report clients can raise questions about any published figure or tool output. Our team responds within one business day during standard office hours.
Our Approach
Why Structured Compute References Matter for AI Planning
Engineering teams working through AI infrastructure decisions face a consistent information problem. Published figures from hardware and cloud vendors are typically measured under conditions that favour the product being sold. Independent academic benchmarks, while methodologically rigorous, often cover workloads or configurations that do not map cleanly onto commercial deployment scenarios. The gap between what teams can find and what they actually need for planning tends to be filled with guesswork or with informal estimates that are difficult to defend in cross-functional discussions.
Nymbus was built to close that gap. Our Benchmark Reference Library compiles figures from documented sources and presents them alongside interpretation notes that explain what each number means in practice. The Comparison Toolkit takes that data and makes it easy for teams to frame trade-offs side by side, without switching between spreadsheets and browser tabs. For organisations preparing a specific internal briefing, the Custom Insight Report brings both into a single written resource tailored to the questions at hand.
Petaling Jaya is home to a growing number of organisations running or planning AI-related workloads, from financial services firms evaluating GPU provisioning to logistics companies exploring edge inference deployments. Nymbus serves that community by providing the kind of calm, well-documented reference materials that make technical planning conversations more productive — and more defensible when they reach senior stakeholders.
Ready to Work with Nymbus?
Reach out to discuss which product fits your current planning stage and we will follow up within one business day.
Get in Touch