NVIDIA· Professional
NVIDIA-Certified Professional: Generative AI LLMs
NCP-GENL
Intermediate-level certification validating ability to design, train, and fine-tune cutting-edge LLMs, applying advanced distributed training techniques and optimization strategies to deliver high-performance AI solutions in production environments.
420+
Questions
7
Pruebas de Práctica
22
Hands-on labs
Lifetime
Updates
What's included
Practice
7 Exam-realistic practice tests
420+ questions
Practice exam · NCP-GENLQ 1 / 65
Correct · explanation
Hands-on
Labs· 22
# vLLM · 10 concurrent
$ asyncio.gather(…)
wall time ....... 3.8s
throughput ...... 1,842 tok/s
intermediate45m
Deploy & Serve LLMs in Production (Jupyter)
Unlock# lora · r=16 · alpha=32
$ trainer.train()
trainable ....... 0.8%
eval_loss ....... 1.24
perplexity ...... 3.45
intermediate45m
Fine-Tune an LLM with LoRA and QLoRA (Jupyter)
Unlockadvanced45m
CUDA Programming Fundamentals
Unlockintermediate40m
Quantize & Optimize LLMs with bitsandbytes
Unlock# torch.profiler · top ops
$ table(sort=cuda_time)
gemm ........ 41%
memcpy ...... 22%
softmax ..... 9%
intermediate35m
Profile PyTorch Training with the Built-in Profiler
Unlockadvanced40m
Advanced RAG: Hybrid Search + Cross-Encoder Reranking
Unlockadvanced45m
Continued Pre-Training: Adapt a Pretrained LM to a New Domain
Unlockintermediate30m
NVIDIA DALI: GPU-Accelerated Data Pipelines
Unlock# dedup + tokenize
$ filter_lang("en")
raw ............ 52k → 18k
tokenizer ...... BPE 32k
train_ready .... 16,420
intermediate45m
Data Preparation for LLM Training
Unlockadvanced55m
RLHF & DPO Alignment
Unlockintermediate45m
Evaluation & Benchmarking LLMs
Unlockintermediate40m
Inference Serving Patterns: Dynamic Batching, Throughput, and the Triton Mental Model
Unlockintermediate35m
Nsight Systems Profiling: Finding the Bottleneck That Costs You 40% of Your GPU
Unlockintermediate40m
Batch Size & Precision Sweep: Finding Your Sweet Spot
Unlockintermediate45m
Retrieval-Augmented Generation (RAG) Pipeline with Local Models
Unlock# determinism flags
$ torch.use_deterministic_…
seed ............ 42
cudnn.benchmark . False
reproducible .... TRUE
intermediate40m
Reproducible Training: The Flags, The Cost, The Artifacts
Unlock
GPU Prointermediate45m
Fine-Tune Stable Diffusion with LoRA: Custom Text-to-Image
Unlock
GPU Prointermediate40m
Synthetic Data Generation for Model Training
Unlock
GPU Proadvanced55m
Train a Small Language Model from Scratch
Unlock
GPU Proadvanced50m
Build a Transformer from Scratch: Attention, Masking & LayerNorm
Unlock
GPU Proadvanced55m
vLLM Production Serving: PagedAttention, Continuous Batching, Prefix Caching
Unlock
Hosted Prointermediate35m
Build NeMo Guardrails for an AI Agent: Jailbreak & Topical Rails
UnlockReady to start preparing?
Unlock every resource above with one purchase. Lifetime access, 30-day refund.
Ready to Pass Your Certification?
Get the practice tests on their own, or go Pro to add 22 hands-on labs for this cert plus every other track on Preporato.
Just Practice Tests
$19.99One-time payment
Lifetime access — pay once, study forever
- 7 full-length practice tests
- 420+ exam-style questions
- Detailed explanations for every answer
- Exam mode & learning mode
- ×No hands-on labs
Best value
Preporato Pro
$24.17/month
Billed yearly — $290/year. Cancel anytime.
- 22 hands-on labs for NCP-GENL + 53 more AI/ML labs
- All practice tests — 31 certifications
- GPU sandboxes + hosted environments
- Flashcards + study guides + articles
- Cancel anytime, no contract
Secure checkout via Stripe
Garantía de Devolución de 30 Días
Instant Access
Lifetime Updates