AI RESEARCH ENGINEER

In this kickoff session, get ready to meet your instructor and dive into the course. We'll introduce you to the objectives, cover essential AI models and technologies, and spark your curiosity with an interactive case study, setting the stage for an engaging learning journey. Get your questions ready for a Q&A session – it's the perfect way to start.

• Instructor introduction
• Course objectives & flow
• AI overview
• Case study: Peeking into the AI Technology Stack
• Q&A session

Let’s get the full picture of the current AI landscape. You’ll explore the historical context and branches of AI systems while delving into the ethical responsibilities that come with them. Real-world case studies, including ProPublica's "Machine Bias" and the Princeton Dialogues on AI Ethics' "Automated Healthcare App," will shed light on recent AI advancements and the ethical challenges they pose. No need for pre-reading; we'll navigate all the complexities together.

• Current AI research landscape overview
• AI systems: historical context
• The different branches of AI systems
• Ethical responsibilities in AI
• Case study: Ethical challenges & AI 

Class 2 is your gateway to the AI toolkit! Discover how these tools can amplify your AI projects, explore their practical applications, and get a sneak peek with live demos. It's a hands-on journey to elevate your understanding and usage of essential AI resources.

• AI research tools & frameworks: intro
• Practical application of AI research tools
• Utilizing AI frameworks in research projects
• Demo: AI tools & frameworks in action

Dive deeper into the latest trends, demystifying the systems and AI production challenges that lie behind them. Explore emerging technologies, and get hands-on with demos and case studies, discovering how industry leaders like Microsoft integrate AI into their frameworks. Plus, engage in a collaborative workshop to whiteboard AI pipelines for real-world use cases.

• AI research trends: significance
• Emerging AI technologies & integration into AI research
• Demo: AI models in emerging tech
• Case study: Microsoft & Copilot
• Workshop: AI pipeline for an emerging technology

In Class 4, we're rolling up our sleeves and getting hands-on, learning the ropes of data analysis and feature engineering. Understand the importance of data in AI research, master advanced analysis techniques for informed decisions, and explore innovative ways to sculpt top-notch data for your AI models.

• Data in AI research trends
• Advanced data analysis & data-based decision-making
• Innovative data representation & feature engineering methods
• Workshop: Handling small & large datasets

Assignment #1: Analyzing & Visualizing Complex Datasets
Select a dataset from our list or bring your own for extra credit. We'll guide you with sample problem statements to shape your dataset's focus. Identify and document two data issues, clean and engineer features for your problem statement, and apply class techniques to visualize your data. Your submission should feature a clear problem statement, a concise slide deck (2-5 slides) detailing data issues, cleanup steps, and key visuals showcasing your dataset's highlights.

Explore and apply state-of-the-art models for computer-vision applications, from object classification to activity recognition. Practice implementing advanced models – gear up to explore the forefront of computer vision research and turn theory into practice in our workshop session.

• Computer vision research: latest trends
• Object classification, detection, image segmentation, and activity recognition
• Transfer learning & fine-tuning: pre-trained models
• Workshop: Implementing a computer vision research solution

Assignment #2: Computer Vision Research Problem
Choose your data and decide whether to custom-train a model or load a pre-trained one, with the option to fine-tune it. Your task is to ensure your AI computer vision model successfully runs end-to-end, producing outputs for classification, detection, segmentation, or activity recognition.

Elevate your teamwork and project planning! Class 6 is all about mastering the collaborative side of AI, especially when working with teams within a business and the broader AI industry. We'll dive into essential tools like Weights and Biases & Jupyter Notebooks, and simulate a real-world project with a computer vision challenge.

• Collaborative tools for AI research
• Effective project planning
• Workshop: Collaboration & PM: computer vision problem

Let’s dissect techniques for Natural Language Processing tasks, with a special emphasis on sentiment analysis. From the basics to cutting-edge models like Transformers, we'll explore it all, topped with a hands-on workshop to analyze real-world text data.

• NLP fundamentals
• Demo: Advanced NLP models
• Workshop: Sentiment analysis & text data

Assignment #3: Sentiment Analysis/Text Classification
Choose between two options: either dive into sentiment analysis on a provided dataset or tackle a text classification task using advanced NLP techniques covered in class. Your work might be featured in class for a constructive discussion on your approach.

What are the best transfer learning and fine-tuning techniques for creating custom text-generation models? We're going to be exploring advanced ML models and getting ready for a workshop to fine-tune and experiment with the power of Generative AI.

• Reviewing assignments: strengths & weaknesses
• Data for ML models: intro
• Advanced ML models: GPT, BERT, RNNs
• Transfer learning & fine-tuning: pre-trained models
• Workshop: Generative AI models 

Assignment #4: Custom Generative AI Text-Model via Fine-tuning
In this project, you'll take a pre-trained NLP model and tailor it to a specific task. Think data analysis, feature engineering, transformer-based models, and fine-tuning techniques. You can use pre-cleaned data or craft your own dataset. Your submission? A concise dataset description, a snippet of code for analysis, your model's feature engineering, your fine-tuned model, and a quick demo of how it runs.

Understand the theory and practice of Generative Adversarial Networks, a foundational and pivotal technique that spun off the domain of image generation. From transforming sketches into photographs to dissecting the nuances behind image generation, we’ll cover the powerful practical applications of GANs in the world of AI.

• Demo & Case study: Image-to-image translation with GANs
• Generating images using GANs: theory
• Workshop: Image generation with GANs

In Class 10, we’ll look into advanced image-generation Generative AI models and techniques (e.g., Stable Diffusion, Latent Consistency Models). By the end of the session, you’ll understand the architecture behind these models and how to utilize them to perform transfer learning effectively.

• Demo & Case study: Shopify’s AI Product Genie
• Generating images: Stable Diffusion models & more
• Workshop: Image generation model & diffusion techniques

Assignment #5: Fine-tuning Diffusion Models for Image Generation
Once again, you'll fine-tune a pre-trained NLP model to craft a specialized model, now – for a specific task in image generation. Apply your skills in data analysis, feature engineering, and transformer-based models learned in previous classes. You can use pre-processed data from dataset hubs or create your own. Your submission should include a brief dataset overview, code-based analysis, feature engineering, model fine-tuning, and a demonstration of model inference.

How do you apply reinforcement learning techniques for robotics and other applications? Understand the unique challenges that may entail. Dive into the essentials, from the basics of reinforcement learning to advanced algorithms. Unpack Google's real efforts, exploring how RL propels robotics into the future in real-world applications.

• Reinforcement learning: intro
• Deep reinforcement learning algorithms
• Workshop: Implementing RL algorithms
• RL algorithms for LLMs fine-tuning
• Case study: Google

In Class 12, discover how to interpret and apply techniques to AI models across various domains like computer vision and robotics. Join us for a workshop, where you'll implement and evaluate XAI techniques, giving you the keys to unlock the solutions to specific challenges.

• Importance of XAI in AI research
• Methods for interpreting & explaining AI models
• Implementing & evaluating XAI techniques
• XAI, computer vision, NLP & robotics
• Workshop & Demo: Implementing XAI for computer vision and robotics

Assignment #6: Reinforcement Learning & Robotics
Apply an RL algorithm to a robotic task using the Gymnasium-Robotics library (refer to the documentation at farama.org). Then, enhance the environment with XAI.

This session is all about AI research with a focus on advanced experimental design and methodology. From A/B testing to causal inference, learn techniques that go beyond the basics, with insights from a guest speaker and a case study in causality and counterfactual analysis for Economics. Get ready to present your findings with confidence!

• Advanced experimental design for AI research
• A/B testing, causal inference, and counterfactual analysis
• Guest speaker
• Case study: Causality & counterfactual analysis
• Preparing & presenting research findings

Assignment #7: A/B Testing
Conduct an A/B test, analyze the results, and prepare a research-oriented report or presentation (2-3 slides).

Ethical AI is vital to get right. Let’s dissect real-world challenges with a case study from the Markkula Center for Applied Ethics based on IBM’s dataset, explore technical solutions to identify and mitigate bias, and learn how to implement ethical guardrails for AI models.

• Case study: IBM’s “Diversity in Faces” dataset
• Techniques to identify & mitigate bias
• Ethical considerations in AI research & deployment
• Demo: Practical implementation with pseudocode
• Workshop: Implementing technical guardrails for an AI model in pseudocode

What are the key steps to success and career advancement in AI? How do you boost your technical expertise and expand your network? Find out in the final class. Discover the ins and outs of advancing in AI and gain industry insights that'll turbocharge your career journey.

• Growth strategy
• Building a professional AI research network
• Conferences, workshops, and meetups
• Networking: social media & online forums
• Guest speaker

Final Project: AI Research Showcase
Throughout the course, you will be bringing your learnings to life in the final project! Tackle a chosen problem statement. Your project involves problem formulation, data analysis, model development, evaluation, and ethical considerations. Present your results in a well-documented report deck, using provided templates and frameworks. Feel free to explore different domains like computer vision, natural language processing, text/image generation, or reinforcement learning.