AI-assisted labs · NSS College of Engineering, Palakkad · KTU

Engineering labs with an AI tutor, on NVIDIA NIM.

Three free, browser-based labs for the S5 Artificial Intelligence (PECST522) and S7 Compiler Lab (CSL411) courses. Every answer is checked by the lab itself; when a student is stuck, an AI tutor running on NVIDIA Nemotron via build.nvidia.com explains the next step. Start with a two-minute video, then pick a lab.

Usage so far / live

1,727
Lab page views
364
AI tutor requests on NVIDIA NIM
94
Student NVIDIA keys tested
3
Labs with an NVIDIA AI tutor

Counted since 1 Sep 2026 from this server's own access logs and the labs' request ledger, refreshed daily (13 Oct 2026). Known bots and scripts are excluded. A whole classroom on the college Wi-Fi shares one IP address, so device counts undercount students. Raw figures: nvidia-labs-stats.json. Independent check: see Independent analytics below.

Independent analytics / Matomo

Measured by Matomo, the open-source analytics platform, self-hosted on this server. The dashboard is public and read-only: no login needed.

447
Lab visits since 1 Sep
8 min 54 s
Average time per visit
86%
Visits from India
58%
On phones or tablets
Sep 2026 256Oct 2026 191

Open the live analytics dashboard ↗ Figures refreshed hourly from Matomo's public API.

Lab pages are those under /ai/ and /compiler-lab/. Visits before 13 Oct were imported from the server's access logs; since 13 Oct every lab page also reports live to Matomo (cookieless, IP-anonymised). Visits from a whole classroom on one college Wi-Fi IP can merge, so these figures undercount students.

Start here / 01

  1. Open build.nvidia.com and sign in. It is free and needs no card.
  2. Open any model page, choose Get API Key, and copy the key that starts with nvapi-.
  3. Pick a lab below, open its AI tutor panel, paste the key and press Test.

Your key stays in your browser. Each tutor request is relayed once to NVIDIA's hosted endpoint and the key is never stored or logged on this server. No key? Every lab still works: the checkers, hints and visualisations run without it.

Labs with the NVIDIA AI tutor / 02

Companion labs / 03

Same courses, no key needed. These run fully in the browser.

How it is built / 04

MODEL
NVIDIA Nemotron 3 Ultra
nvidia/nemotron-3-ultra-550b-a55b on the OpenAI-compatible endpoint integrate.api.nvidia.com/v1, with an automatic fallback model from the same NVIDIA catalog.
GROUNDING
The lab checks, the model explains
Correctness comes from deterministic checkers (a LEX/YACC runner, a logic verifier, a Wumpus rule engine). The model only sees the student's current step and the checker's verdict, so it explains rather than guesses.
PRIVACY
Bring-your-own key, never stored
Keys live in the student's browser. Student text is not logged. Usage counters are aggregate only: no names, no roll numbers.

Workshops & partnership / 05

Open the AI tutor to every learner.

Today each student must create their own NVIDIA key, which stops many first-time visitors and students from other colleges before they try a lab. We are looking for support from NVIDIA to:

  • offer a shared, rate-limited “try without a key” tutor on all three labs;
  • run hands-on AI and Compiler Lab workshops built on these labs for KTU colleges, alongside NVIDIA Deep Learning Institute courses;
  • extend the same pattern to more KTU lab courses.

Faculty contact: Kiran V K, Assistant Professor, Dept. of CSE, NSS College of Engineering, Palakkad · IEEE CS Top 30 Early Career Professional 2024.

NVIDIA, NIM and Nemotron are trademarks of NVIDIA Corporation. These labs use NVIDIA's public API catalog; this site is independent and is not endorsed by NVIDIA.