AI/ML Training & Internship Program

AI/ML Training & Internship

Home

About The Course

Artificial Intelligence and Machine Learning focus on building systems that can learn from data, recognize patterns, and make intelligent decisions. This course introduces the core concepts of AI and ML, helping students understand how modern applications like recommendation systems, chatbots, and predictive analytics work. With a practical approach, learners gain hands-on experience in applying algorithms to real-world problems and developing data-driven solutions.

Key points

Lessons of the Course

MODULE 1:FOUNDATIONS & CLASSICAL MACHINE LEARNING

✓

Python Essentials

Variables, loops, functions, OOP basics
Library: Python 3.9+
Project: Build a CLI calculator with error handling

✓

Data Manipulation with NumPy & Pandas

Arrays vs DataFrames, filtering, grouping, merging
Libraries: NumPy, Pandas
Project: Clean and analyze a messy sales dataset

✓

Version Control with Git

Commits, branches, GitHub workflows
Tool: Git, GitHub / GitLab
Project: Create portfolio repository with proper structure

✓

Visualization & Exploratory Data Analysis (EDA)

Static Visualization
- Matplotlib: Custom plots, subplots, styling
- Seaborn: Statistical plotting, heatmaps, pairplots
- Libraries: Matplotlib, Seaborn
- Project: Visualize COVID-19 trends with multiple chart types
Interactive Visualization & EDA
- Plotly / Dash basics, correlation analysis, outlier detection
- Libraries: Plotly, Pandas Profiling
- Project: Interactive dashboard for housing data
EDA Case Study
- Complete EDA pipeline on a real-world dataset
- Tool: Jupyter Notebook
- Deliverable: Professional EDA report with insights

MODULE 2: Mathematics & Classical ML Algorithms

✓

Linear Algebra for ML

Vectors, matrices, dot products - implemented in NumPy
Application: How recommendation systems use matrix operations
Library: NumPy for all math implementations

✓

Calculus Essentials

Derivatives, partial derivatives, gradients
Visualization: Gradient descent with animation
Tool: Custom Python visualization

✓

Statistics & Probability

Distributions, Bayes' Theorem, hypothesis testing
Application: Spam filter probability calculation
Library: SciPy Stats

✓

Math in Practice

Implement linear regression from scratch using only NumPy
Compare with the scikit-learn implementation
Project: Build a regression model without ML libraries

✓

Supervised & Unsupervised Learning

Supervised Learning I
- Linear/Logistic Regression, Evaluation Metrics (MSE, Accuracy, Precision/Recall)
- Library: scikit-learn
- Project: Customer churn prediction
Supervised Learning II
- Decision Trees, Random Forest, SVM
- Focus: When to use which algorithm
- Project: Credit scoring system
Unsupervised Learning
- K-Means Clustering, PCA, Hierarchical Clustering
- Library: scikit-learn
- Project: Customer segmentation for marketing
Model Evaluation & Selection
- Cross-validation, hyperparameter tuning (GridSearchCV)
- Bias-Variance tradeoff in practice
- Tool: scikit-learn, Yellowbrick for visualization
- Project: Optimize a model pipeline end-to-end

✓

Project: Real Estate Price Prediction & Recommendation System

Predict house prices (regression) + Recommend similar houses (clustering)
Full pipeline: Data cleaning → EDA → Feature engineering → Multiple models → Evaluation
Deliverable: GitHub repo with code + Streamlit app demo

MODULE 3: DEEP LEARNING & ADVANCED TOPICS

✓

Introduction to Neural Networks

Perceptrons, activation functions, forward/backward propagation
Library: TensorFlow / Keras
Project: Build NN from scratch using NumPy

✓

Building with Keras/TensorFlow

Sequential / Functional API, callbacks, early stopping
Library: TensorFlow 2.x, Keras
Project: MNIST digit classification (98%+ accuracy)

✓

Optimization & Regularization

Dropout, BatchNorm, different optimizers
Visualization: TensorBoard for tracking experiments
Tool: TensorBoard, Weights & Biases (optional)

✓

Hyperparameter Tuning at Scale

Keras Tuner, automated hyperparameter optimization
Library: Keras Tuner
Project: Optimize network architecture automatically

✓

Advanced Neural Architectures

Convolutional Neural Networks (CNNs)
- Conv layers, pooling, modern architectures (ResNet, EfficientNet)
- Project: Cats vs Dogs classifier with data augmentation
Transfer Learning
- Fine-tuning pre-trained models (VGG16, ResNet50)
- Library: TensorFlow Hub, tf.keras.applications
- Project: Medical image classification with limited data
Introduction to NLP & RNNs
- Text preprocessing, word embeddings, simple RNNs
- Library: TensorFlow, NLTK / spaCy basics
- Project: Sentiment analysis on movie reviews
Time Series & Sequence Models
- LSTM, GRU for time series prediction
- Project: Stock price prediction (simplified)

✓

Intelligent Image Recognition System

Multi-class image classification with CNNs
Full deployment: FastAPI + Docker + Hugging Face Spaces
Deliverable: Live web application with model inference

MODULE 4: CAPSTONE & SPECIALIZATION

✓

Structure

Students choose from curated project tracks or bring their own idea.

✓

Available Tracks

Computer Vision: Object detection system with YOLO
NLP: Text classification/chatbot with transformers
Time Series: Forecasting system with LSTM / Prophet
Recommendation System: Collaborative filtering + deep learning

✓

Project Planning & Development

Problem definition, data collection, literature review
MVP development, baseline model implementation
Mentorship: Daily standups, code reviews, and troubleshooting sessions

✓

Advanced Implementation & Optimization

Advanced modeling, hyperparameter optimization
Deployment, documentation, presentation preparation
Tools: Student's choice from learned stack + mentor recommendations

✓

Advanced Topics & Industry Readiness

Transformers & Hugging Face
- BERT, GPT basics, using pre-trained models
- Library: Transformers, Hugging Face Hub
- Project: Fine-tune BERT for a custom task
Advanced Visualization
- Interactive dashboards with Streamlit / Gradio
- Library: Streamlit, Gradio
- Project: Build ML web app in 2 hours
MLOps Tools
- DVC for data versioning, MLflow pipelines
- Tools: DVC, MLflow
- Project: Reproducible ML pipeline
Production Considerations
- Model quantization, ONNX format, edge deployment
- Library: ONNX Runtime, TensorFlow Lite
- Demo: Deploy model to mobile
Ethics in AI
- Bias detection, fairness metrics, explainable AI (SHAP / LIME)
- Library: SHAP, Fairlearn
- Case Study: Analyze model bias in hiring dataset

✓

Portfolio Development & Final Presentations

Portfolio polishing, GitHub READMEs, LinkedIn optimization
Final project presentations to industry panel
Graduation, certifications, next steps guidance

TOOLS & LIBRARIES SUMMARY

✓

Core Stack

Programming: Python 3.9+
ML/DL: scikit-learn, TensorFlow 2.x / Keras, PyTorch (optional)
Data: NumPy, Pandas, SQL basics
Visualization: Matplotlib, Seaborn, Plotly, Streamlit
Deployment: FastAPI, Docker, Hugging Face Spaces

✓

Project Management

GitHub for code hosting
Notion / Trello for project tracking
Slack / Discord for communication

✓

Assessment Structure

Weekly Projects (40%) - Hands-on implementation
Mid-term Project (25%) - End-to-end ML system
Capstone Project (35%) - Industry-level solution with documentation

✓

Differentiators of This Program

Zero to Deployment in 12 weeks
Industry Tools First - Teach what companies actually use
Project Portfolio - 8+ completed projects for GitHub
Deployment Focus - Every major model gets deployed

What Our Students Say

Man standing outdoors in a green mountainous landscape, wearing a blue sleeveless jacket and jeans.

Learning MERN Stack at Excendra was transformative. Practical projects, clear explanations, and industry-focused mentorship helped me develop strong full-stack skills and confidence to build professional applications.

Mehran Muzaffar Software Developer COMTECH GROUP

I completed Full Stack Development training at Excendra. Practical projects and clear teaching built my skills and confidence. With their support, I secured a job soon after the course and recommend Excendra for tech career.

Hurain Javaid Software Engineer Sun-Net Inc.