Let's Ship It - with AWS! ML Edition
Are you a developer working in a fast-paced startup environment looking to understand how to use machine learning (ML) to solve real business problems? This free interactive Twitch training is designed to help you apply ML to the most common real-life use cases affecting startups—even if you have limited access to data or resources. Over the course of 8 episodes, you’ll learn hands-on how to build ML models for a wide variety of scenarios: to predict demand, measure churn, personalize your offerings, and more.
If you have a developer background or similar and are looking to develop ML skills you can use to solve real-world problems, this show is the perfect place to start. Join us at twitch.tv/aws.
Sign up to get replays delivered to your inbox after the episodes air, plus links to resources from each episode.
Explore all of our on-demand shows to see what’s coming up live, and to find on-demand episodes from past series for AWS Training and Certification.
Episode 1: Machine Learning for Startup Developers 101
Welcome to “Let’s Ship It – with AWS! ML Edition”! In our first episode, we’ll explore machine learning (ML) concepts: what is ML, which problems you can solve with ML, and different model types. By the end of this episode, you’ll be able to identify common use cases for ML for startups that you can pick up—even with limited resources, data, and no ML background.
Instructor: Aaron Hunter, Fred Graichen
Episode 2: Phases of ML Workflows and Tooling
In machine learning (ML), you "teach" a computer to make predictions, or inferences, through a step-by-step process. In this episode, you’ll explore phases of ML workflows and best practices for startups. You’ll learn steps to have a working ML solution: from problem framing and data preparation to training and deployment. We’ll also cover what tooling can make this process efficient.
Instructor: Aaron Hunter, Fred Graichen
Episode 3: How to Perform Sentiment Analysis on Customer Feedback
At a startup, reacting quickly to customer feedback is critical. How do you gather insights from large amounts of data as you keep growing your customer base? You’ll be fine-tuning a pre-trained model using Amazon SageMaker and the Amazon Reviews Polarity dataset, which consists of around 35 million reviews from Amazon, and classify the reviews into either positive or negative feedback.
Instructor: Aaron Hunter, Fred Graichen
Episode 4: Give Personalized Recommendations
Join us as we stream LIVE from Las Vegas during our re:MARS event! Did you know that personalization can reduce customer acquisition costs by as much as 50 percent and lift revenues by 5 to 15 percent (McKinsey Research)? In this episode, we explore how you can use personalization to lower your startup’s customer acquisition costs. We use an example of an online retailer that sells unique all-occasion gifts to show you how you can build great recommendation systems that learn from past data and predict interests to offer your customers tailored experiences.
Instructor: Aaron Hunter, Fred Graichen
Episode 5: Generating Value from Understanding the Visual World
From e-Commerce to Health Care, startups are increasingly building applications that are enhanced with predictions from computer vision models. Computer vision is a sub-field of machine learning on how computers see and understand the world visually. In this episode you will learn how to build a model that makes predictions from visuals, using the example of computer vision for medical imaging: your model will be able to predict if an image of cells contains cancer cells.
Instructor: Aaron Hunter, Fred Graichen
Episode 6: Predicting Customer Churn
For startups it’s crucial to get to market early with a minimum viable product and iterate on it as you gather feedback. For this reason, reducing churn - the number of customers that a business is losing - is one of the most important goals of startups looking for hypergrowth. In this episode, you’ll learn how you can use machine learning to automatically identify unhappy customers and to evaluate incentive optimization strategies to retain revenue. We use an example of churn that is familiar to all of us: leaving a mobile phone operator.
Instructor: Aaron Hunter, Fred Graichen
Episode 7: Accelerate Your Growth
In a highly startup competitive market, the most efficient growth strategies can be the key difference to achieve unicorn status. Whether it's e-mail, mail, or phone, we want to only target prospective customers that are likely to engage with a specific offer to keep customer acquisition costs low. In this episode, we’ll build a machine learning (ML) model to predict whether a customer will make a buying decision based on readily available information using Amazon SageMaker Autopilot - an AutoML capability to select the best ML model without writing a line of code.
Instructor: Aaron Hunter, Fred Graichen
Episode 8: Startup ML Best Practices and Next Steps + Guest Speaker Expert Q&A
Get your questions answered in this Q&A session with a surprise guest speaker in the machine learning (ML) space for our last episode. You’ll learn more about best practices for ML, common challenges across startups, trend predictions and what you can do next in your ML journey.
Instructor: Aaron Hunter, Fred Graichen, Special Guest (TBA)