AI Full Stack Training in Hyderabad at Version IT

• Python setup, IDEs, environment (VS Code, Jupyter)
• Variables, data types, operators
• Input/output handling
• Conditional statements & loops
• Data structures:
o Lists, Tuples, Sets, Dictionaries
• Functions:
o Arguments, return types, recurssion
• File handling (CSV, JSON, TXT)
• Exception handling & debugging
• Object-Oriented Programming:
o Classes, objects, inheritance, polymorphism
o Decorators, generators, iterators
o Shallow vs Deep Copy
• Introduction to NumPy & Pandas

• What is AI, ML, Deep Learning
• Real-world AI applications
• Introduction to Generative AI
• Types of GenAI:
o Text, Image, Audio, Code generation
• AI lifecycle & architecture
• APIs & model usage concepts

• Text preprocessing:
o Tokenization, stemming, lemmatization
o Stopwords removal
• Feature engineering:
o Bag of Words
o TF-IDF
• Word embeddings:
o Word2Vec (CBOW & Skip-gram)
o Sentence embeddings
• Hands-on NLP pipelines

• Neural Networks basics
• RNN (Recurrent Neural Networks)
• LSTM & GRU (limitations solved)
• Bidirectional RNN
• Encoder–Decoder architecture
• Sequence-to-Sequence models

• Why Attention is needed
• Attention mechanism (intuition + math)
• Transformer architecture:
o Embeddings & positional encoding
o Multi-head attention
o Residual connections & normalization
o Feedforward layers
• Training pipeline of transformers

• Evolution of LLMs
• Model types:
o Encoder-only (BERT)
o Decoder-only (GPT)
o Encoder-Decoder (T5)
• Tokens & tokenization
• Parameters:
o Temperature, Top-p, max tokens
• Context window concept

• OpenAI API usage
• Prompt-based interaction
• Structured outputs
• Function calling
• API integration in Python apps

• Prompt design fundamentals
• Zero-shot vs Few-shot prompting
• Chain-of-Thought (CoT)
• Role-based prompting
• Output formatting techniques
• Prompt optimization strategies

• LangChain architecture
• Components:
o LLMs, Prompts, Chains
• Memory:
o Buffer, summary, window memory
• Output parsers
• Runnables & pipelines
• Tool integration

• Document loaders:
o PDF, TXT, Web APIs
• Text splitting strategies
• Embedding generation
• Vector databases:
o FAISS
o Pinecone
o Chroma
• Similarity search concepts

• RAG architecture & workflow
• Retriever design
• Naïve RAG implementation
• Advanced RAG:
o Multi-query retrieval
o Reranking
• RAG evaluation techniques
• Real-world use cases (chatbots, Q&A systems)

• What are AI agents
• Tool calling & APIs
• ReAct framework (Reason + Act)
• Multi-agent systems
• Agent communication
• Workflow orchestration

• LangChain vs LangGraph
• Workflow types:
o Sequential
o Parallel
o Iterative
• Stateful agents
• Streaming responses
• Chatbot with memory & tools

• RAG + Agents integration
• Multi-agent orchestration
• Guardrails & safe AI
• AI for automation (resume screening, analytics)
• No-code AI tools (LangFlow)

• Fine-tuning basics
• Instruction tuning
• Quantization:
o GGML vs GGUF
• Performance vs cost trade-offs

• Metrics:
o BLEU, ROUGE, METEOR
• Evaluation pipelines
• Experiment tracking
• Debugging LLM outputs

• CI/CD for AI apps
• GitHub Actions
• Model versioning
• Dataset management
• Experiment tracking

• Streamlit app deployment
• API deployment
• Cloud basics
• Scaling LLM apps

• Git basics:
o Repositories, commits, branching
• Docker:
o Containers & images
o Deployment workflows