This course covers the tools and techniques commonly utilized for production machine learning in industry. Students learn how to provide web interfaces for training machine learning or deep learning models with Flask and Docker. Students will deploy models in the cloud through Amazon Web Services, gather and process data from the web, and display information for consumption in advanced web applications using Plotly and D3.js. The students use PySpark to make querying even the largest data stores manageable.

Explain why students should care to learn the material presented in this class.

Weeks to Completion: 7
Total Seat Hours: 37.5 hours
Total Out-of-Class Hours: 75 hours
Total Hours: 112.5 hours
Units: 3 units
Delivery Method: Residential
Class Sessions: 14 classes, 7 labs

• DS 2.1

Students by the end of the course will be able to:

1. Implement Machine Learning or Deep Learning on a Web App using Flask and Flask-Restplus
2. Dockerize the Flask Web App and develop it on Amazon Web Services (AWS)
3. Read large datasets from S3 through Boto
4. Implement Advanced Visualizations using a Chart.js/D3.js frontend, and a Python Backend
5. Work on Big Data using Pyspark, H2O and Pandas

Course Dates: Monday, August 26 – Wednesday, October 9, 2019 (7 weeks)

Class Times: Monday and Wednesday at 3:30–5:20pm (13 class sessions)

HW1: Deploying ML/DL model to AWS without docker
HW2: API on AWS that logs user interactions when working with the ML/DL model. Create a MongoDB or SQL Database. We want each user interactions with Flask-Keras API will be written into DB. Including the image file name, the prediction result and date for Mnist application.
HW3: For Titanic dataset, do Pyspark to get the same result as we got with Pandas

• Coming Soon!

All projects will require a minimum of 10 commits, and must take place throughout the entirety of the course

• Good Example: 40+ commits throughout the length of the course, looking for a healthy spattering of commits each week (such as 3-5 per day).
• Bad Example: 10 commits on one day during the course and no others. Students who do this will be at severe risk of not passing the class.
• Unacceptable Example: 2 commits the day before a project is due. Students who do this should not expect to pass the class.

We want to encourage best practices that you will see working as a professional software engineer. Breaking up a project by doing a large amount of commits helps engineers in the following ways:

• It's much easier to retrace your steps if you break your project/product/code up into smaller pieces
• It helps with being able to comprehend the larger problem, and also will help with your debugging (i.e. finding exactly when you pushed that piece of broken code)
• It allows for more streamlined, iterative communication in your team, as it's much easier to hand off a small change to someone (updating a function) than a huge one (changed the architecture of the project)

Through this requirement, we hope to encourage you to think about projects with an iterative, modular mindset. Doing so will allow you to break projects down into smaller milestones that come together to make your fully-realized solution.

• Passing the exam is a requirement for passing the class.
• You will have 2 hours to complete this exam - it will be in class using paper and pencil, or a format of the instructor's choosing
• There are no retakes of the exam.
• If you have a disability that needs an accommodation such as extended time or a different format, please take advantage of our accommodations program.

To pass this course you must meet the following requirements:

• Complete all required tutorials
• Pass all projects according to the associated project rubric
• Pass the final summative assessment >=75%
• Actively participate in class and abide by the attendance policy
• Make up all classwork from all absences

Just like any job, attendance at Make School is required and a key component of your success. Attendance is being onsite from 9:30 to 5:30 each day, attending all scheduled sessions including classes, huddles, coaching and school meetings, and working in the study labs when not in a scheduled session. Working onsite allows you to learn with your peers, have access to support from TAs, instructors and others, and is vital to your learning.

Attendance requirements for scheduled sessions are:

• No more than two no call no shows per term in any scheduled session.
• No more than four excused absences per term in any scheduled session.

Failure to meet these requirements will result in a PIP (Participation Improvement Plan). Failure to improve after the PIP is cause for not being allowed to continue at Make School.

Academic Honesty
Accommodations for Students
Attendance Policy
Diversity and Inclusion Policy
Grading System
Title IX Policy
Program Learning Outcomes