Artificial Intelligence Lab

Designed for absolute beginners with no prior technology background. This course equips you with practical, day-one usable AI skills using today's most powerful tools — helping you work smarter in any profession, business or field of study.

• Fundamentals of Artificial Intelligence & how it is transforming industries, education and daily life
• Hands-on usage of ChatGPT (GPT-4o), Google Gemini & GitHub Copilot for real tasks
• Prompt engineering: crafting precise, context-rich prompts to get accurate AI outputs
• AI tools for automated writing, report generation, email drafting & content creation
• Image generation with DALL-E 3 & Canva AI for creative & marketing work
• Applying AI in business workflows: marketing automation, HR screening & customer support bots
• Understanding AI limitations, hallucinations & fact-checking strategies
• Responsible AI: data privacy, copyright concerns, ethical use guidelines & AI governance basics

A practical, project-based introduction to Python programming with a direct focus on AI and Machine Learning applications. No prior coding experience required. By the end, students will write Python programs, work with data and build basic ML models.

• Python basics: variables, data types, strings, lists, dictionaries, loops & functions
• Object-Oriented Programming (OOP): classes, objects, inheritance & encapsulation
• File handling: reading & writing CSV, JSON & text files
• NumPy: arrays, matrix operations, broadcasting & numerical computing
• Pandas: DataFrames, loading datasets, cleaning, filtering, grouping & merging data
• Data visualisation: line charts, bar plots, scatter plots & heatmaps using Matplotlib & Seaborn
• Introduction to Machine Learning with Scikit-learn: classification, regression & clustering algorithms
• Mini project: end-to-end data analysis & ML model on a real-world dataset

Industry-level, hands-on training in Computer Vision and AI deployment infrastructure using MPIT's NVIDIA DGX H200 GPU cluster. Students will build, train and deploy real-world vision AI systems — from image processing to production-grade deployment pipelines.

• Image processing fundamentals using OpenCV: filtering, edge detection, morphology & colour spaces
• Video analysis: frame extraction, optical flow & real-time video processing pipelines
• Object detection & tracking: YOLO (v8/v9) training on custom datasets with annotation tools
• Deep Learning for vision: CNN architectures — ResNet, VGG, EfficientNet & Transfer Learning
• Image segmentation: semantic vs instance segmentation (Mask R-CNN, SAM)
• NVIDIA CUDA programming basics: GPU memory management & accelerated model training on H200 cluster
• Containerisation with Docker: building, tagging & deploying AI model containers
• Scalable deployment using Kubernetes: pods, services, load balancing & monitoring
• REST API for vision models using FastAPI: building inference endpoints
• End-to-end CV pipeline: data collection → annotation → training → evaluation → deployment

Master the full Generative AI engineering stack — from understanding and fine-tuning large language models to building autonomous multi-agent AI systems. Hands-on training on MPIT's NVIDIA H200 GPU cluster prepares students for the most in-demand AI engineering roles.

• Generative AI landscape: GPT family, LLaMA, Mistral, Gemma & open-source models
• Transformer architecture: attention mechanism, tokenisation & model inference
• Fine-tuning LLMs: full fine-tuning vs PEFT (LoRA, QLoRA) using Hugging Face Transformers
• Retrieval-Augmented Generation (RAG): vector embeddings, indexing & semantic search
• Vector databases: Pinecone, ChromaDB, Weaviate — setup, ingestion & querying
• LangChain framework: chains, memory, agents, tools & LangGraph for stateful workflows
• LlamaIndex for document Q&A, knowledge base & enterprise RAG pipelines
• Agentic AI systems: multi-agent orchestration using AutoGen & CrewAI
• Prompt engineering for production: few-shot, chain-of-thought, structured outputs
• Deploying GenAI applications: REST APIs (FastAPI), Streamlit front-ends & cloud hosting

The most comprehensive AI engineering programme at MPIT CoE. Combining Computer Vision & AI Infrastructure (40 hrs) and GenAI, LLMs, RAG & Agentic AI (40 hrs) into a single immersive 80-hour programme — with capstone project and placement support.

• OpenCV image & video processing, YOLO object detection on custom datasets
• CNN architectures, Transfer Learning, image segmentation (Mask R-CNN, SAM)
• NVIDIA CUDA basics, GPU-accelerated training on NVIDIA DGX H200 cluster
• Docker containerisation, Kubernetes deployment & FastAPI inference endpoints

• LLM fine-tuning with LoRA/QLoRA on Hugging Face Transformers
• RAG pipeline design with vector databases (Pinecone, ChromaDB)
• LangChain, LlamaIndex & LangGraph for LLM-powered applications
• Multi-agent systems with AutoGen & CrewAI
• Deploying GenAI apps with FastAPI & Streamlit

• Capstone project: build & deploy a production-ready AI product combining CV + GenAI on NVIDIA H200
• Live project demos & peer review sessions
• Industry mentorship sessions with practicing AI engineers
• GitHub portfolio documentation & project write-up
• Placement assistance & referrals to hiring partners