6 Months Program
Generative AI Developer
Master LLMs, PEFT, RLHF, and Production Deployment - From Transformer Internals to Gen-AI MLOps
6 Months
Duration
6 hrs/week
Weekly Coaching
30-50
Batch Size
$699
Course Fee
Learning Path
LLM Fundamentals
→
Large-Language Models in Production
→
Gen-AI Dev Path
Course Structure
Months 1-2
Transformer Internals & Prompt Design
Learning Outcomes
- ✓Understand the transformer architecture powering LLMs
- ✓Master different types of prompt engineering strategies
- ✓Gain hands-on experience with inference and API-based LLMs
Weekly Topics
- • Transformer building blocks: attention, embeddings, positional encoding
- • Tokenization & vocabulary handling
- • Prompt engineering: zero-shot, few-shot, chain-of-thought
- • Model APIs: OpenAI, Anthropic, Hugging Face Inference
- • Basics of context management & output formatting
Practical Work
- • Implement a transformer attention mechanism from scratch (PyTorch)
- • Design prompt templates for multiple NLP tasks
- • Build a prompt benchmarking script
Assessment
- • Weekly quizzes on theory + coding
- • GitHub repo submission of the prompt engineering library
Months 3-4
PEFT (LoRA / Soft Prompts) & RLHF
Learning Outcomes
- ✓Fine-tune LLMs efficiently using Parameter-Efficient Fine-Tuning
- ✓Apply LoRA, adapters, and soft prompts to custom datasets
- ✓Implement Reinforcement Learning with Human Feedback (RLHF)
Weekly Topics
- • Introduction to PEFT & why full fine-tuning is costly
- • LoRA theory & implementation in Hugging Face PEFT library
- • Soft prompt tuning for domain adaptation
- • RLHF pipeline: reward model, policy optimization
- • Data labeling strategies for RLHF
Practical Work
- • Fine-tune a 7B model with LoRA on domain-specific data
- • Implement a soft-prompt tuned chatbot
- • RLHF lab with a custom reward model
Assessment
- • Repo review of LoRA-tuned model
- • Mid-term oral defense: PEFT vs. full fine-tuning trade-offs
Months 5-6
Gen-AI MLOps, Evaluation & Cost Control
Learning Outcomes
- ✓Deploy LLMs in production with scalability and monitoring
- ✓Build evaluation frameworks for generative tasks
- ✓Optimize inference cost and latency
Weekly Topics
- • Containerizing LLM inference with Docker
- • Managed services: Azure OpenAI, Vertex AI, AWS Bedrock
- • Evaluation metrics for LLM output: BLEU, ROUGE, human eval
- • Token cost analysis & batching strategies
- • Model quantization & optimization
Practical Work
- • Deploy a LoRA-tuned model using a managed service
- • Build a LangSmith/Weights & Biases evaluation dashboard
- • Implement token usage monitoring with alerting
Assessment
- • Final project: End-to-end generative AI application
- • Deployment link + GitHub repo + recorded demo presentation
Deliverables for Learners
Projects
- •Transformer attention module from scratch
- •LoRA fine-tuned LLM
- •RLHF-trained chatbot
- •Managed-service deployed generative AI app
Certifications
End-to-end badge for course completion with comprehensive assessment of all modules
Resources
- •Annotated code labs
- •Recorded sessions
- •Project templates
Ready to Master Generative AI?
Join our comprehensive 6-month program and become a certified Generative AI Developer with hands-on expertise in LLMs, PEFT, RLHF, and production deployment.