About Aegis-GPU

Aegis-GPU is a free, independent study platform for NVIDIA certifications. It started as a hands-on GPU operations simulator for NCA-AIIO and is growing one certification at a time.

Independence

Aegis-GPU is not affiliated with, endorsed by, or sponsored by NVIDIA. We have no access to exam content. Every practice question is original, written from NVIDIA's public blueprints and documentation, and cites the documents it relies on. If you think an item is wrong, use the feedback button in the simulator; corrections are prioritised.

How content is verified

• Certification facts (duration, question count, price, domains, weights, objectives) are parsed from each official NVIDIA certification page by tools/refresh_cert_catalog.py, which also reports drift. Fields that could not be confirmed are shown as unverified. Last full check: 2026-10-08.
• GPU Xid errors come from NVIDIA's Xid Catalog (Ampere and newer). One generated file feeds the API, the tutor and the Xid reference, and a test fails if they disagree.
• Full courses (NCA-AIIO and NCA-GENL) hold a stricter standard: every question, guide point, glossary term and lab is tagged to an official objective and stores a verbatim quote from an NVIDIA-owned source. tools/verify_evidence.py re-fetches every cited page and checks each quote; a CI test fails if anything lacks an objective or an NVIDIA source, or if question counts drift from the official domain weights.
• Practice questions are mapped to an official objective and list their sources. Questions are reviewed against NVIDIA documentation; some v1 items were corrected in v2 (for example DCGM ECC field IDs 310/311 and Xid 48/63/64/79 guidance).
• Ask Aegis answers from this content. When an AI model is enabled it is instructed to answer only from retrieved passages and cite them; otherwise you get a rule-based answer built from the same passages.

Privacy

• No third-party analytics, ads, trackers or cookies. Traffic is counted from our own web-server logs (page views and unique IPs per day) and the logs rotate weekly.
• Without an account, your practice history, bookmarks and plan live only in your browser's local storage.
• With an account we store your username, a bcrypt password hash, a hashed recovery code, your learner state, feedback you send, and a log of questions you ask the tutor. No email address is required.
• When AI tutor answers are enabled, your question and the retrieved course passages are sent to the configured model provider to write the answer. Do not paste secrets into the tutor.
• You can delete your account and all associated data from the account page.

Open source

Source code: github.com/NdumLab/aegis-GPU.