AI & GenAI
Executive Program in AI Engineering & Agentic AI
Self-paced GenAI & LLM engineering: LLMs, prompting, RAG, AI agents, fine-tuning and deployment.
What you'll learn
- Build with LLMs: prompting, structured outputs and function calling
- Design and ship production RAG applications
- Build AI agents that use tools and take actions
- Fine-tune and customise models when it is the right choice
- Deploy, monitor and cost-control GenAI in production
Skills & tools you'll gain
Prompt EngineeringRAGVector DatabasesAI AgentsFunction CallingFine-tuning (LoRA/PEFT)PyTorchLangChainLlamaIndexFastAPIDockerLLMOps & DeploymentGuardrails & Monitoring
Course content 36 modules · 266 topics · 37h 36m total
▸Module 1: Introduction to AI Engineering & the GenAI LandscapeUnderstand what an AI Engineer does today and how generative AI reshaped the field.10 topics · 1h 23m
▶What is an AI Engineer?Preview5:37▶AI vs ML vs Deep Learning vs GenAIPreview6:14
🔒The Modern AI Stack6:51
🔒Where LLMs Fit In7:28
🔒Capabilities & Limits of GenAI8:05
🔒Common AI Application Patterns8:42
🔒Responsible & Ethical AI9:19
🔒The AI Product Lifecycle9:56
🔒Roles in an AI Team10m
🔒Industry Use Cases11m
▸Module 2: Development Environment & Working with LLM APIsSet up a professional workspace and make your first model calls.10 topics · 1h 22m
▶Python Environment & VirtualenvPreview11m
🔒Notebooks vs Scripts5:24
🔒Managing API Keys & Secrets6:01
🔒Anatomy of an LLM API Call6:38
🔒Requests, Responses & Tokens7:15
🔒Temperature & Sampling7:52
🔒Streaming Responses8:29
🔒Rate Limits & Error Handling9:06
🔒Cost Awareness9:43
🔒Examples: OpenAI, Claude, Open Endpoints10m
▸Module 3: Python for AI EngineersSharpen the Python patterns you will use every day on the job.9 topics · 1h 11m
🔒Functions, Modules & Packages10m
🔒Working with JSON11m
🔒Type Hints & Dataclasses5:11
🔒Async Basics for API Calls5:48
🔒Environment Variables6:25
🔒Logging7:02
🔒Reading & Writing Files7:39
🔒Calling REST APIs8:16
🔒Clean Code for AI Apps8:53
▸Module 4: Data Handling for AILoad, clean and explore data the way every AI project requires.9 topics · 1h 19m
🔒NumPy Arrays & Vectors9:30
🔒Pandas DataFrames10m
🔒Loading CSV / JSON / Parquet10m
🔒Cleaning & Missing Values11m
🔒Filtering & Aggregation11m
🔒Basic Visualization5:35
🔒Train / Test Concepts6:12
🔒Working with Text Data6:49
🔒Data Quality for AI7:26
▸Module 5: Machine Learning Foundations (Just Enough)The core ML ideas an AI engineer needs, without the heavy math.8 topics · 1h 14m
🔒Supervised vs Unsupervised8:03
🔒Features & Labels8:40
🔒Classification vs Regression9:17
🔒Training vs Inference9:54
🔒Overfitting & Generalization10m
🔒Accuracy, Precision & Recall11m
🔒The Bias-Variance Idea11m
🔒Why GenAI Still Needs ML Thinking5:22
▸Module 6: Embeddings & Vector RepresentationsUnderstand how meaning becomes math, the backbone of modern AI.8 topics · 1h 5m
🔒What Are Embeddings?5:59
🔒Turning Text into Vectors6:36
🔒Semantic Similarity7:13
🔒Cosine Similarity & Distance7:50
🔒Embedding Models (concept)8:27
🔒Visualizing Embeddings9:04
🔒Search, Clustering & Recommendations9:41
🔒Limitations of Embeddings10m
▸Module 7: Neural Networks FundamentalsOpen the black box, how neural networks actually learn.8 topics · 1h 2m
🔒Neurons, Weights & Biases10m
🔒Layers & Activation Functions11m
🔒Forward Propagation5:09
🔒Loss Functions5:46
🔒Gradient Descent (intuition)6:23
🔒Back-Propagation (intuition)7:00
🔒Epochs, Batches & Learning Rate7:37
🔒Why Deep Networks Work8:14
▸Module 8: Deep Learning with PyTorchBuild and train a neural network in the industry-standard framework.9 topics · 1h 20m
🔒Why PyTorch Is the Default8:51
🔒Tensors & Operations9:28
🔒Autograd10m
🔒Defining a Model (nn.Module)10m
🔒The Training Loop11m
🔒Optimizers & Schedulers11m
🔒GPU vs CPU5:33
🔒Saving & Loading Models6:10
🔒Datasets & DataLoaders6:47
▸Module 9: Training, Evaluation & Avoiding OverfittingGet models that generalize, not just memorize.7 topics · 1h 4m
🔒Train / Validation / Test Splits7:24
🔒Regularization & Dropout8:01
🔒Early Stopping8:38
🔒Hyperparameter Tuning9:15
🔒Reading Loss Curves9:52
🔒Transfer Learning (intro)10m
🔒Model Evaluation in Practice11m
▸Module 10: NLP FoundationsHow machines process human language, the road to LLMs.7 topics · 52m
🔒What Is NLP?11m
🔒Text Preprocessing5:20
🔒Tokenization Basics5:57
🔒Bag-of-Words & TF-IDF6:34
🔒Sequence Modeling Problems7:11
🔒Classic vs Modern NLP7:48
🔒Classification, NER, Summarization8:25
▸Module 11: Tokenization & Word EmbeddingsHow text is broken down and represented inside language models.7 topics · 1h 2m
🔒Tokens vs Words vs Characters9:02
🔒Subword Tokenization (BPE concept)9:39
🔒Vocabulary & Token IDs10m
🔒Word Embeddings (Word2Vec idea)10m
🔒Contextual Embeddings11m
🔒Token Limits & Context5:07
🔒Counting Tokens for Cost5:44
▸Module 12: The Transformer ArchitectureThe breakthrough that powers every modern LLM, explained clearly.7 topics · 57m
🔒Why Transformers Replaced RNNs6:21
🔒Self-Attention (intuition)6:58
🔒Multi-Head Attention7:35
🔒Positional Encoding8:12
🔒Encoder vs Decoder8:49
🔒Pre-training vs Fine-tuning9:26
🔒From Transformer to GPT-style Models10m
▸Module 13: How Large Language Models WorkA clear mental model of what LLMs are doing under the hood.8 topics · 1h 7m
🔒Next-Token Prediction10m
🔒Pre-training at Scale11m
🔒What "Parameters" Mean11m
🔒Emergent Capabilities5:31
🔒Strengths & Failure Modes6:08
🔒Hallucinations Explained6:45
🔒Determinism & Randomness7:22
🔒Why Prompts Matter7:59
▸Module 14: Model Families: Open vs APINavigate the LLM ecosystem and choose the right model for the job.7 topics · 1h 6m
🔒API vs Open-Weight Models8:36
🔒Model Sizes & Trade-offs9:13
🔒Latency, Cost & Quality9:50
🔒Running Open Models Locally (concept)10m
🔒Multimodal Models11m
🔒How to Choose a Model11m
🔒Examples: GPT, Claude, Llama, Mistral5:18
▸Module 15: Context, Tokens & LimitationsWork within the real constraints of production LLMs.7 topics · 54m
🔒Context Windows Explained5:55
🔒Input vs Output Tokens6:32
🔒Managing Long Inputs7:09
🔒Truncation & Summarization7:46
🔒System / User / Assistant Roles8:23
🔒Conversation Memory9:00
🔒Token Budgeting & Cost Control9:37
▸Module 16: Prompting Fundamentals & PatternsGet reliable, high-quality output from any LLM.8 topics · 1h 4m
🔒Anatomy of a Good Prompt10m
🔒Zero-shot vs Few-shot10m
🔒Role & System Prompts11m
🔒Chain-of-Thought Prompting5:05
🔒Instruction Clarity5:42
🔒Delimiters & Formatting6:19
🔒Common Prompting Pitfalls6:56
🔒Reusable Prompt Templates7:33
▸Module 17: Structured Outputs & Function SchemasMake LLM output machine-readable and production-ready.7 topics · 1h 10m
🔒Why Structured Output Matters8:10
🔒Asking for JSON8:47
🔒JSON Schema & Validation9:24
🔒Function / Tool Schemas10m
🔒Parsing & Error Recovery10m
🔒Constrained Generation (concept)11m
🔒Output Reliability Techniques11m
▸Module 18: Prompt Evaluation & IterationTreat prompts like engineering artifacts, test and improve them.7 topics · 51m
🔒What Makes a Prompt "Good"?5:29
🔒Building a Test Set6:06
🔒Manual vs Automated Evaluation6:43
🔒LLM-as-a-Judge (concept)7:20
🔒Versioning Prompts7:57
🔒A/B Testing Prompts8:34
🔒Reducing Hallucinations9:11
▸Module 19: Introduction to RAGGive LLMs access to your own knowledge, the most in-demand AI skill.7 topics · 1h 0m
🔒Why RAG? Limits of LLM Knowledge9:48
🔒The RAG Pipeline Overview10m
🔒Retrieve, Augment, Generate11m
🔒Grounding & Citations11m
🔒RAG vs Fine-tuning5:16
🔒RAG Architecture Patterns5:53
🔒Real-World RAG Use Cases6:30
▸Module 20: Embeddings & Vector DatabasesStore and search knowledge by meaning at scale.7 topics · 1h 2m
🔒From Documents to Embeddings7:07
🔒What Is a Vector Database?7:44
🔒Indexing & Similarity Search8:21
🔒Metadata Filtering8:58
🔒Hosted vs Self-Hosted9:35
🔒Choosing a Vector Store10m
🔒Examples: Pinecone, Chroma, pgvector, FAISS10m
▸Module 21: Chunking & Retrieval StrategiesThe unglamorous decisions that make or break a RAG app.7 topics · 50m
🔒Document Loading & Parsing11m
🔒Chunk Size & Overlap5:03
🔒Chunking Strategies5:40
🔒Embedding the Chunks6:17
🔒Top-K Retrieval6:54
🔒Relevance & Recall7:31
🔒Handling Tables, PDFs & Mixed Content8:08
▸Module 22: Building a Production RAG ApplicationTurn a RAG prototype into something you can ship.7 topics · 1h 7m
🔒End-to-End RAG Architecture8:45
🔒Prompt Assembly with Context9:22
🔒Source Citations in Answers9:59
🔒Handling "I do not know"10m
🔒Conversation + Retrieval11m
🔒Caching11m
🔒Putting a UI on RAG5:27
▸Module 23: Advanced Retrieval TechniquesPush retrieval quality from okay to excellent.7 topics · 55m
🔒Hybrid Search (keyword + vector)6:04
🔒Re-ranking Results6:41
🔒Query Rewriting & Expansion7:18
🔒Filtered Retrieval7:55
🔒Multi-Query Retrieval8:32
🔒Hierarchical / Parent-Child Chunks9:09
🔒Reducing Irrelevant Context9:46
▸Module 24: RAG Evaluation & QualityMeasure and improve what your RAG system actually returns.7 topics · 57m
🔒Retrieval Metrics (recall, precision)10m
🔒Answer Faithfulness11m
🔒Groundedness & Citations11m
🔒Building a RAG Test Set5:14
🔒Detecting Hallucinations5:51
🔒Continuous Evaluation6:28
🔒Common RAG Failure Modes7:05
▸Module 25: Introduction to AI AgentsMove from single answers to systems that take actions.7 topics · 1h 6m
🔒What Is an AI Agent?7:42
🔒LLM + Tools + Memory + Loop8:19
🔒Agents vs Workflows8:56
🔒The ReAct Pattern (reason + act)9:33
🔒Planning & Decomposition10m
🔒When Not to Use Agents10m
🔒Agent Use Cases11m
▸Module 26: Tool Use & Function CallingLet an LLM call your code, APIs and the outside world.7 topics · 48m
🔒Function / Tool Calling Explained5:01
🔒Defining Tools & Schemas5:38
🔒Parsing Tool Calls6:15
🔒Executing & Returning Results6:52
🔒Multi-Tool Selection7:29
🔒Error Handling in Tool Use8:06
🔒Safety & Guardrails for Actions8:43
▸Module 27: Agent Frameworks & OrchestrationBuild multi-step agents with the patterns the industry uses.7 topics · 1h 4m
🔒Orchestration Frameworks (concept)9:20
🔒Agent Memory & State9:57
🔒Multi-Step Reasoning10m
🔒Multi-Agent Patterns11m
🔒Human-in-the-Loop11m
🔒Observability & Tracing5:25
🔒Examples: LangChain, LlamaIndex6:02
▸Module 28: When & Why to Fine-TuneKnow the right tool, fine-tuning, RAG or better prompting.7 topics · 59m
🔒Fine-tuning vs RAG vs Prompting6:39
🔒What Fine-tuning Can & Cannot Do7:16
🔒Cost & Effort Trade-offs7:53
🔒Use Cases for Fine-tuning8:30
🔒Data Requirements9:07
🔒Risks & Pitfalls9:44
🔒A Decision Framework10m
▸Module 29: Fine-Tuning TechniquesUnderstand modern, efficient ways to customize a model.7 topics · 54m
🔒Full vs Parameter-Efficient Tuning10m
🔒LoRA & PEFT (concept)11m
🔒Instruction Tuning5:12
🔒Quantization Basics5:49
🔒Hardware Considerations6:26
🔒Hosted Fine-tuning Services7:03
🔒Open-Model Fine-tuning (concept)7:40
▸Module 30: Datasets & Evaluation for Fine-TuningGarbage in, garbage out, build datasets that actually help.7 topics · 1h 10m
🔒Collecting & Curating Data8:17
🔒Formatting Training Examples8:54
🔒Cleaning & De-duplication9:31
🔒Train / Validation Split10m
🔒Evaluating a Fine-tuned Model10m
🔒Avoiding Catastrophic Forgetting11m
🔒Iterating Safely11m
▸Module 31: Serving AI ApplicationsWrap your AI in something users and other systems can call.7 topics · 52m
🔒Building an API with FastAPI5:36
🔒Request / Response Design6:13
🔒Streaming to Clients6:50
🔒Building UIs (Streamlit / Gradio)7:27
🔒Authentication Basics8:04
🔒Async & Concurrency8:41
🔒Connecting Front-end to AI Back-end9:18
▸Module 32: Containerization & Cloud DeploymentShip your app reliably, anywhere.7 topics · 1h 1m
🔒Why Containers? Docker Basics9:55
🔒Building a Container Image10m
🔒Secrets in Production11m
🔒Deploying to the Cloud (concept)11m
🔒Scaling Basics5:23
🔒CI/CD Overview6:00
🔒Free Hosting Options for Demos6:37
▸Module 33: Cost, Latency, Guardrails & MonitoringRun AI in production without surprises.7 topics · 1h 3m
🔒Token Cost Optimization7:14
🔒Caching & Batching7:51
🔒Latency Reduction8:28
🔒Input / Output Guardrails9:05
🔒Prompt-Injection Awareness9:42
🔒Logging & Observability10m
🔒Monitoring Quality & Privacy10m
▸Module 34: Capstone Planning & BuildScope and build a real, portfolio-worthy GenAI product.7 topics · 51m
🔒Choosing a Capstone Problem11m
🔒Designing the AI Architecture5:10
🔒RAG / Agent / Fine-tuning Choice5:47
🔒Building the Core Pipeline6:24
🔒Iterating with Evaluation7:01
🔒Documentation & README7:38
🔒Adding Guardrails8:15
▸Module 35: Deployment & DemoGet your capstone live and demo-ready.5 topics · 50m
🔒Deploying the Capstone Publicly8:52
🔒Building the Demo UI9:29
🔒Performance & Cost Check10m
🔒Recording a Demo Video10m
🔒Writing the Case Study11m
▸Module 36: Portfolio & Career ShowcasePackage everything to get hired.6 topics · 45m
🔒GitHub Portfolio Setup11m
🔒Showcasing Your Projects5:34
🔒Writing About Your Work6:11
🔒Interview Prep for AI Roles6:48
🔒Building Your Personal Brand7:25
🔒The CareerByteCode Showcase8:02
About this course
A self-paced BytePlay edition of the CareerByteCode AI Engineer program. 36 production-focused modules from LLM foundations to deployment, with hands-on projects. Learn at your own pace, then take the exam, build realtime projects on ByteLabs and start earning. (DEMO course, safe to remove.)
Requirements
- Python basics (functions, loops, packages)
- A laptop and curiosity
- No prior AI or ML experience needed
Who this course is for
- Engineers moving into AI / GenAI roles
- Students building a hireable AI portfolio
- Professionals shipping LLM features at work
Your trainer
View all self-paced courses by Demo Trainer →🏆
Finished the videos? Go further
Turn what you learned into a certificate, a portfolio and income. Each one is its own next step.