Open Source Models
LLaMA, Mistral, Qwen, DeepSeek - running locally, fine-tuning with LoRA and QLoRA, quantization, evaluation, and production deployment with vLLM.
7 Modules. From Download to Production.
Own your models. No API costs. No rate limits. Full control.
Model Ecosystem
LLaMA, Mistral, Qwen, DeepSeek, Phi - the open-source landscape, architectures, and licensing.
What you'll master
- LLaMA Family Architecture
- Mistral and Mixtral Architecture
- Qwen, DeepSeek, and International Models
- Phi and Small Language Models
- Code and Math Specialized Models
- Multimodal Open-Source Models
- Model Licensing and Compliance
- HuggingFace Hub and Model Cards
8 lessons
Start for Free →
Running Locally
llama.cpp, Ollama, MLX, LM Studio - running 70B models on consumer hardware, GGUF, and benchmarking.
What you'll master
- llama.cpp and GGUF Format
- Ollama and Local Model Management
- MLX for Apple Silicon
- LM Studio and GUI Tools
- Docker and Containerized Local Inference
- Hardware Requirements and Selection
- Privacy and Air-Gapped Deployment
- Benchmarking Local Model Performance
8 lessons
Start for Free →
LoRA and QLoRA Fine-Tuning
LoRA mathematics, QLoRA 4-bit training, Axolotl and TRL frameworks, and model merging.
What you'll master
- LoRA Mathematics and Implementation
- QLoRA 4-Bit Fine-Tuning
- Selecting Target Modules and Rank
- Training Data Preparation
- Axolotl and TRL Training Frameworks
- Evaluating Fine-Tuned Models
- Advanced PEFT Methods
- Merging and Model Soup Techniques
8 lessons
Start for Free →
Quantization in Practice
GPTQ, AWQ, PTQ, QAT - quantizing open-source models, quality vs speed tradeoffs, and production deployment.
What you'll master
- Post-Training Quantization Methods
- GPTQ In Depth
- AWQ In Depth
- Quantization-Aware Training
- Quantization Benchmarking
- Quantization for Vision Models
- Quantization Error Debugging
- Deploying Quantized Models in Production
8 lessons
Start for Free →
Fine-Tuning Pipelines
Instruction tuning, RLHF, DPO, continual learning, synthetic data, and fine-tuning economics.
What you'll master
- Full Fine-Tuning vs PEFT
- Instruction Tuning at Scale
- RLHF and DPO for Open Models
- Continual Learning and Domain Adaptation
- Synthetic Data and Self-Improvement
- Fine-Tuning Hyperparameter Search
- Monitoring and Debugging Training
- Fine-Tuning Cost and ROI Analysis
8 lessons
Start for Free →
Evaluating Open Models
Open LLM leaderboards, safety evaluation, hallucination testing, and building custom eval harnesses.
What you'll master
- Open LLM Leaderboard and Benchmarks
- Task-Specific Evaluation Design
- Safety and Bias Evaluation
- Factuality and Hallucination Evaluation
- Code Generation Evaluation
- Reasoning and Math Evaluation
- Long Context Evaluation
- Building an Evaluation Harness
8 lessons
Start for Free →
Production Deployment
vLLM, TGI, Kubernetes auto-scaling, load balancing, monitoring, and multi-model serving.
What you'll master
- vLLM Architecture and Deployment
- TGI and Alternatives
- Kubernetes and Auto-Scaling for LLMs
- Load Balancing and Request Routing
- Monitoring LLM Services
- Rate Limiting and Cost Control
- Model Versioning and Canary Releases
- Multi-Model Serving Architecture
8 lessons
Start for Free →
Run your own models. Own your AI stack.
From GGUF on a MacBook to 70B vLLM clusters - every step covered.
Start Learning Free →