1. Re-cap on various Python libraries (Pandas, numpy etc.)
2. Text Processing using ML – Embeddings
3. Hands on Coding for generting Embeddings and doing clustering
4. Review of NN and NN helps generate Embeddings – Word2Vec, Glove
5. Hands on Coding for using Embeddings for similarity of words
6. Issues with NN based Embeddings
7. CNN based NLP tasks
8. Use CNN to solve Sentiment Analyzer
9. Limitations of NN, CNN in NLP
10. Sequential Data (and Data dependence)
11. RNN for processing Seq Data
12. Diff RNN Architecture (deep dive into Math)
13. RNN architecture for different problems
14. Forward Propagation in RNN
15. Back propagation in RNN
16. Vanishing and Exploding Gradient
17. What it means to have Vanishing and Exploding gradients in terms of Learning
18. LSTM, BiDirectional LSTM, Deep BiDirectional RNN
19. Intro to Keras – Checkpointing, LearningRate scheduling etc.
20. Hands on coding of Sentiment Analyzer using RNN, LSTM BiDir etc
21. Encoder Decoder Architecture
22. Issues with RNN based Encoder Decoder Architecture

RNN+Attention
Issues with RNN+Attention
Attention without RNN aka Transformers
Transfer Learning in CNN and now with Transformers
Consume Transformer based architecture
Clustering using the Transformer

What is Gen AI
Gen AI Tools – Chat GPT, ChatGPT Plus, Gemini,….
Gen AI Technologies foundation
Auto Encoder + VAE
Diffusion Models
SLM, MLM, LLM – Transformer based Generative AI
Encoder Based
Decoder Based
Encoder + Decoder based
Difference between Transformer based earlier models and Gen AI models – “Instruction Trained”
Different ways of consuming/using.building Gen AI apps
Gen AI as a UI User (No coding required)
Gen AI as a API user (no ML/DL required, only Python coding)
Gen AI as a designer (ML+DL+NLP needed)
Gen AI as a Creator
Use Gen AI Tools – Chat GPT, Gemini via UI (brief over view, later to be do Prompt Engineering)
Use Gen AI as a API user
Use Gen AI as a Designer – Prompt Engineering vs Model Finetuning
Use Gen AI as a Model Creator
Key LLMs and how to choose a LLM for a project
Open Sopurced Foundational LLMs
Closed Source LLMs
Gen AI- LLM Vocabulary – Context, Prompt, Completions
Re-cap on various Python libraries (Pandas, numpy etc.)
Re-cap on how NLP is used in Generative AI
Prompt Engineering Strategy
Zero Shot, one Shot, Few Shot Learning
What happens in Prompt Engg? Does model get trained?
Key parameters in LLM/Gen AI – temperature, Greedy.Random, Top K, Top P
Prompt Engineering and Art of Effective Prompts
Solve some NLP problems using ChatGPT UI
Solve some NLP problems using ChatGPT API (needs Python coding)
Chunking strategy, Pre-processing – OpenAI, Limitations
Langchain, PAL, COTR
Industry use case and Existing demos for reference

Key Limitations of LLM – Hallucination
How ti appears both in Fine Tuned as well as Prompt Engineered LLM
Ways to address that – RAG
How to index relevant docuemtns – Vecto DB
Key Vector DB – Weaviate, FAISS
Measure relevant docs – Cosine Theta, Euclidean Distance
Insert Documents in Vector DB
Do some Searches on Vector DB
RAG Techiques and patterns, and LLM Evaluation
RAG in Action

Fine tuning
Use case walk-through (Knowledge Search)
Advance topics such as Quantization, Inference Optimization etc.
Responsible AI and LLMOps
How to Deploy OpenAI key in Azure Function