MLOps and Systems

Program Syllabus

About FourthBrain
FourthBrain trains aspiring Machine Learning practitioners in the technical and practical skills
necessary to contribute immediately to an AI team. This program prepares students with the
tools and skills to deploy, test, and monitor ML models and pipelines for timely inferencing in
production environments. We infuse values of collaboration, communication, empathy, and
equity throughout the program.

Program Outcomes
At the end of the program, students will be able to:

● Build and validate well known ML/DL model prototypes on a variety of ML use cases such
as Data Streams (eCommerce), Natural Language Processing and Computer Vision
● Apply transfer learning for ML Model deployment on cloud platforms
● Understand the impact of data drift and concept drift in ML pipelines
● Apply AutoML and collaborative frameworks such as MLflow
● Build and maintain CI/CD pipelines for cloud-based ML-Model Deployments
● Create rapid prototypes from state-of-the-art pre-trained models
● Perform pre-production testing using A/B patterns, shadow patterns and Canary patterns
● Apply Apache Spark for big data workflows
● Deploy ML models using optimized Kuberenetes workflows
● Apply production-specific software services using Prometheus, grafana cloud, elk stack
(AWS), fluentD and cloudwatch (AWS)
● Integrate themselves into industrial teams as key contributors for production-ready
environments

Our program emphasizes more than just technical skills. In addition to the outcomes listed
above, students will also be able to:

● Communicate effectively to technical and non-technical audiences

● Approach their career goals with the skills and knowledge of how to apply their newly
gained skills in their chosen field

fourthbrain.ai
Updated 4/15/21
 Weekly Schedule
June 2021 Chort

The MLOps and System program is a 12-week program that includes both individual and team
projects.

Week Topics Project

Orientation Introduction to the program

6/19 ● Review syllabus, weekly
assignments, and expectations for
individual and team projects

Week 1 Statistical ML Models for Data Streams E-commerce use-case for data
6/26 ● Classification, Regression, wrangling, compression, meta-data
Unsupervised learning, tagging, metrics and visualization
● AutoML: autosklearn, TPOT,
● Linear and non-Linear Models:
linear regression, Random Forest,
SVM, kNN, k-means, logistic
regression
● Visualization using D3, Tableau.

Break Independence Day

Week 2 Deep Learning Models for Computer Computer vision use case for data
7/10 Vision use cases pre-processing, transfer learning,
● Classification metrics and reporting.
● Object Detection
● Semantic segmentation
● Metrics and Transfer Learning
● RAPIDS GPU

Week 3 ML API Development and Deployment on Rapid Prototyping for NLP use case
7/17 GCP and AWS
● Flask
● FastApi
● Tensorflow serving
● Tensorflow lite for optimization
latency

fourthbrain.ai
Updated 4/15/21
Week 4 Big Data Processing Recommendation engine use case:
7/24 ● Intro to Apache Spark Collaborative filtering use case for
● Intro to Hadoop product rating
● Pyspark (MLlib)
● Collaborative filtering Capstone Project Finalized
● Introduction to Databases (relational
vs no-SQL)
● Databricks

Week 5 Cloud-based APIs Computer Vision use case:

7/31 ● Intro to Kubeflow Hyper-parameterization using
● Collaborative platforms: MLflow MLflow
● AutoML in GCP, Azure, AWS
● Hyper parameter optimization

Week 6 Midterm Project

8/7 Capstone Project Proposal

Week 7 ML Deployment Deployment use case

8/14 ● Docker Containers
● Kubernetes
● FluentD
● Elk Stack

Week 8 MLOps Pipelines Deployment use Case

8/21 ● CI/CD Pipelines
● Gitlab, Jenkins
● YAML/XML
● Profiler

Week 9 Application Production Testing on Google Deployment use Case

8/28 Optimize
● Canary pattern
● A/B Pattern
● Shadow Pattern

Break Labor Day

Week 10 Production Software Build and Analytics Deployment use case

9/11 ● Prometheus
● Grafana Cloud
● Cloudwatch
Special Topics

fourthbrain.ai
Updated 4/15/21
 ● Building models from scratch
● Hyper-parameterization
● optimization
● quantization

Week 11 Final Demo Day Capstone Project Demo

9/18

Capstone Project
You’ll develop a capstone project with one or two other classmates. Your project will be designed
to demonstrate your understanding of ML Pipeline and infrastructure requirements; data
collection and preprocessing; deployment and testing; system design; outcomes; and
extendability.

Sample Team Project 1: There is a pre-existing application that is used for online
shopping. The Search engine searches based on text entries. Your goal is to create a new
function “Find similar” for a product using its image, and to implement the updated ML
model pipeline.

Sample Team Project 2: There is an existing application that detects faces in camera
images. Your goal is to extend the application to face detection/identification for faces
with masks, and to implement the updated ML model pipeline.

Communication
Communication with technical and non-technical colleagues is a crucial skill for all engineers. We
emphasize the importance of regular verbal and written communication throughout the program.
You’ll regularly collaborate with your peers in breakout sessions to reinforce engineering team
settings. For your capstone project, you’ll collaborate with your project team over a period of
several weeks. Your team will regularly update the cohort on your status in both verbal and
written form. The final project deliverable will also include a presentation and technical report.

Career Growth
The MLOps and Systems program is designed for you to acquire the skills and knowledge
required to work on an MLOps team. Whatever your reason for taking the program - to get a new
role at a new company, to gain skills for your current job, or just for fun - we will support your

fourthbrain.ai
Updated 4/15/21
career growth by helping you connect to professionals and employers, via guest speaking
events and inviting employers to the final project presentation day.

The following is a list of sample roles that candidates who successfully complete the MLOps and
Systems program are expected to qualify for:

● MLOps Engineer
● ML Tech Manager
● Junior ML Engineer
● QA Engineer / Test Engineer
● ML Systems Integration
● Solutions Engineer
● DevOps on ML/AI Teams

Career services assistance is available after graduation to help ensure that all candidates
achieve their career goals.

fourthbrain.ai
Updated 4/15/21