Course

Fine-Tuning & Model Customization

When and how to customize LLMs — prompting vs. RAG vs. fine-tuning, and the practical techniques (LoRA, QLoRA, SFT, DPO) — without the classic mistake of using fine-tuning to add facts.

Module 1 freeAdvanced ~5h

Start free — The customization spectrum

Fine-tuning is powerful and widely misunderstood. The most common mistake is reaching for it to add knowledge — which is RAG's job — when fine-tuning is really for shaping behavior, style, and format. This course teaches ML engineers and technical practitioners the decision framework (prompting → RAG → fine-tuning), the real techniques (LoRA, QLoRA, SFT, RLHF vs. DPO), the tooling, and the pitfalls (catastrophic forgetting, safety degradation).

It is grounded in the primary research and current practice, and honest about what fine-tuning can and can't do. Current for 2026; specific model and tool details change fast, so the course emphasizes durable concepts over version numbers.

What you'll be able to do

Decide when to prompt, use RAG, or fine-tune
Understand and apply LoRA, QLoRA, SFT, and preference optimization (DPO)
Prepare data and evaluate fine-tuned models properly
Avoid the pitfalls — catastrophic forgetting, overfitting, and safety degradation

Fine-tuningLoRA / QLoRARAG vs. fine-tuningModel evaluationLLM customization

Curriculum

Module 1: Foundations

Free

The customization spectrum, when to fine-tune, fine-tuning vs. RAG, and the cost/tradeoffs.

Module 2: Fine-Tuning Techniques

Premium

Full vs. PEFT, LoRA and QLoRA, instruction tuning/SFT, and preference optimization (RLHF, DPO).

Full fine-tuning vs. parameter-efficient fine-tuning12 min
LoRA and QLoRA15 min
Supervised fine-tuning and instruction tuning12 min
Preference optimization: RLHF and DPO15 min

Module 3: Data, Tooling & Evaluation

Premium

Data quality and formatting, the tooling ecosystem, quantization, and evaluating fine-tuned models.

Data quality and formatting15 min
The fine-tuning tooling ecosystem12 min
Quantization for efficient models12 min
Evaluating fine-tuned models12 min

Module 4: Pitfalls & Responsible Practice

Premium

Catastrophic forgetting, safety degradation, when smaller fine-tuned models win, and the road ahead.

Catastrophic forgetting and overfitting12 min
Fine-tuning and safety alignment12 min
When a smaller fine-tuned model wins12 min
The road ahead12 min

Get the full course — one-time $29.

Start Module 1 free today. Buy once for lifetime access to the remaining 3 modules — no subscription.

One-time payment · lifetime access

Hand-written & fact-checked Certificate on completion Module 1 free

Want all 53? Get the All-Access Bundle for $99 — one purchase, every course.

Student reviews

Reviews come from students who own this course. Enroll to share yours.

Loading reviews…