AI Engineering on Cloud and AIOps

1. Understanding Supervised Machine Learning, Unsupervised Machine Learning and Reinforcement Learning
2. Data Manipulation for Machine Learning
3. Exploratory Data Analysis for Machine Learning
4. Essential Mathematics for Machine Learning and AIDefinition of Generative AI
5. Generative AI Applications
6. Definition of LLMs

1. Regression Supervised Machine Learning Linear RegressionDecision Tree RegressorRandom Forest RegressorOverfitting, Underfitting and Regularization‍
2. Classification Supervised Machine LearningLogistic RegressionDecision Tree ClassifierRandom Forest Classifier‍
3. Unsupervised Machine LearningKMeans ClusteringHierarchical Clustering Dimensionality ReductionAnomaly detection‍
4. Understanding Reinforcement learning: Basic concepts(Action, reward, process), MPD, RL, DRL, DQN, etc.‍
5. MLOps for Machine Learning model deployment

1. Exploring Deep Learning
2. Understanding Neural Networks with TensorFlow
3. Deep dive into Neural Networks with TensorFlow
4. Artificial Neural Networks (ANN)
5. Convolutional Neural Networks (CNN)
6. Object detection using YOLO framework
7. Transfer Learning in CNN
8. Recurrent Neural Networks (RNN) and LSTMs

1. Architechtures of RNN - One to One, Many to Many, Many to One etc
2. Understanding Text for NLP
3. Tokenization, Stemming and Lemmatization, Stopwords and Keywords in NLP
4. Text Vectorization using TF-IDF, Bag of Words and Word 2 Vec
5. Use cases across domains in NLP

1. How Generative AI works
2. Text generation
3. Image generation
4. Audio generation and video generation
5. Using Hugging face to access models for text generation, image generation
6. GANs
7. Types of GANs - ProGAN, SRGan, CycleGAN
8. Auto Encoders, Variational Auto Encoders (VAEs), Diffusion Models

1. Definition of LLMs
2. LLM Use Cases
3. Prompt Tuning
4. Attention Mechanism
5. Transformer Model and architecture
6. Encoder-Decoder arrangements
7. Train and Generate text using Encoder-decoder Architecture
8. BERT for Transfer learning
9. Leveraging multiple pretrained LLMs from Hugging Face
10. Fine tuning LLMs

1. Introduction to Prompt Engineering
2. Successful and Unsuccessful prompts
3. Types of Prompting
4. Introduction to Open AI, GPT , Open AI Playground
5. Cost & latency considerations when calling APIs (OpenAI, Azure, AWS).
6. Multimodal prompting for GPT 4
7. Image generation using Open AI DALLE 3
8. Prompt evaluation
9. Implementing Agents and Chains
10. Implementing zero-shot-react, conversational-react agents with LangChain
11. Open AI Function (Tool Integration)
12. Testing various LLMs with Prompt Engineering

1. Introduction to Synthetic Data
2. Generating Synthetic Data
3. Synthetic Data for LLMs
4. Real-world Applications and Use Cases
5. Hands-on generating and using Synthetic Data

1. Understanding AI Embeddings
2. Advanced Retrieval Techniques
3. Hugging Face Embeddings
4. Vector Databases in AI
5. CRUD operations with Vector Databases
6. RAG - Retreival Augmented Generation
7. RAG solutions using Open AI models and Hugging face models
8. Ethical Considerations in AI Embeddings
9. Navigating AI Hallucinations, Drift, and Bias
10. Embeddings in Real-world Applications
11. Embeddings Optimization and Fine-tuning
12. Embeddings Security and Privacy
13. LLM Ops and model deployment best practices

1. Agents, Agentic AI and Multi-Agent Systems
2. Agent Definition & Autonomy
3. Simple vs. Knowledge-Based Agents
4. Reflex vs. Goal-Driven Agents
5. Microsoft AutoGen
6. Agent Architecture (Perception, Decision, Action)
7. Integrating Knowledge Bases (RAG, Domain Data)
8. Measuring Performance (Success Rate, Resource Usage)
9. Hierarchical Agent Planning
10. Multi-Step Reasoning with LLM
11. Memory & Long-Term Context
12. Integrating Retrieval Augmentation in Agent Workflows
13. Domain-Specific Knowledge & Dynamic Prompting

As part of this IIT Cloud Computing Course, you'll gain hands-on knowledge of cloud infrastructure, deployment models, and leading cloud platforms that power modern AI applications.

1. Cloud Ecosystems
2. Introduction to Cloud Ecosystem
3. Definitions
4. Cloud characteristics
5. Deployment models
6. Leading Service providers (AWS, Google, Azure, etc.)
7. Comparing AWS, Azure, and GCP core services for compute, storage, and AI/ML.
8. Data Centres and their components
9. Service (SaaS, IaaS, PaaS)
10. Issues & Challenges

1. Understanding hypervisors
2. Reference model
3. Virtualisation characteristics
4. Principles of hypervisor design interfaces
5. Types of hypervisors (type-1 and type-2)
6. Differences between Type-1 and Type-2 hypervisors.
7. Design methods of hypervisors (full virtualization, para virtualization, and hardware-assisted virtualization)
8. Memory Virtualisation
9. I/O virtualisation
10. OS virtualization
11. Comparative Analysis of hypervisors
12. Understanding performance, requirements, and bottleneck

1. Understanding LLM Deployment Architectures
2. Containerizing LLM Inference Services (e.g., using FastAPI + Docker)
3. Managing GPU Workloads in Kubernetes
4. Scaling LLM APIs with Kubernetes and Istio
5. Optimizing Latency and Throughput for LLM Containers
6. Secure Access and Rate Limiting for AI APIs
7. CI/CD for LLM-Powered Microservices
8. Monitoring and Logging for LLM Containers
9. Model Versioning and Rollbacks
10. Cost Optimization Strategies for LLM Inference in Production

1. Infrastructure security: Network-level security
2. Host-level security
3. Application-level security
4. Data security and storage: Data privacy and security issues
5. Jurisdictional issues raised by data location
6. Identity and access management
7. Access control
8. IAM, Key Management Services, and zero-trust architecture trust, reputation, risk authentication in cloud computing
9. Client access in the cloud
10. Cloud contracting model
11. Commercial and business considerations

1. EC2 Deep Dive and AMIs
2. EBS vs S3 vs EFS – Storage Solutions
3. Load Balancing and Auto Scaling Basics
4. Intro to Serverless: AWS Lambda
5. Using AWS Bedrock for GenAI (including foundation models)
6. Deploying Open-Source LLMs on EC2/EKS
7. Fine-Tuning and Inference
8. Pipelines on Cloud
9. Cost & Performance
10. Considerations for LLM Workloads
11. SageMaker Pipelines for Model Training and Inference
12. Model Versioning, A/B Testing, and Rollbacks
13. Security and Compliance for GenAI in Production

Total Duration - 132 Hours

Hours trained by IIT - 66 Hours

Hours trained by Futurense - 66 Hours