Machine Learning
Interviews
From beginner to pro
Susan Shu Chang
Susanshu.com
About me
Principal data scientist at Elastic (Elasticsearch)
Built ML in telecom, fintech, social platforms, etc.
6x PyCon speaker and international keynotes
Machine Learning Interviews
Available on
Amazon: https://amzn.to/4aOjO26
O’Reilly:
https://learning.oreilly.com/library/view/machine-
learning-interviews/9781098146535/
Book companion site: Susanshu.substack.com
Warmup question:
Where are you tuning in from?
Week 1:
Introduction to ML roles,
required skills, and resumes
Learning objectives
In this course, you’ll:
● Learn the skill sets required for each type of machine learning role and evaluate
your strengths and weaknesses (Week 1)
● Understand how to craft a relevant ML resume (Week 1)
● Understand the technical fundamentals of machine learning interviews, such as
machine learning theory, statistics, and coding assessments. (Week 2)
● Gain skills for acing behavioral interview questions, and how to shape your
machine learning career narrative. (Week 2)
Week 1 discussion:
What stage are you at in your machine
learning job search?
What’s your biggest goal?
Testing the waters?
Veteran of several interviews?
Just started applying?
Week 1, Part 1
Overview of the Machine Learning Field
The Machine Learning lifecycle
Machine learning Machine learning
Data
development deployment
The Machine Learning lifecycle (detail)
Source: Machine Learning
Interviews by Susan Shu
Chang (O’Reilly)
Week 1, Part 2
Overview of ML roles/jobs
Machine learning career
● What products do you want to build?
● What impact do you want to have?
● What’s the salary?
Confusing job titles; big mistakes
● A candidate with good skills might be applying to the “wrong roles”
with the same job title.
● After looking at the job description carefully, I helped them identify
better roles to apply to.
Data roles
● Data Scientist ● Applied Scientist
○ “Product Data Scientist” ● Data Engineer
● Machine learning engineer ● Data Analyst
● MLOps engineer
Data roles
● Data Scientist ● Applied Scientist
○ “Product Data Scientist” ● Data Engineer
● Machine learning engineer ● Data Analyst
● MLOps engineer
Role confusion is a common frustration for jobseekers
Data Engineer
Source: Malsha Ranawaka
Data Scientist
● Started out as the umbrella term
● New offshoot term: “Product Data Scientist”
● Now, this job title could mean many things
○ Depends on the job description
Machine Learning Engineer
● Infrastructure work
● Model deployment, MLOps
● Model training and development
○ Part of current (newer) trend
Source: ml-ops.org
Applied Scientist
● ML models
Source: Amazon
MLOps Engineer
● Infrastructure work
● Model deployment, MLOps
● Automating ML deployments
More info on my O’Reilly MLOps Superstream keynote
Data Analyst
● Important to business decisions
○ Example: finding average sales of N business line across past X weeks
○ Automating the above via dashboards
● “Product Data Scientist”
● A good entry point for a data career
Previous trend of job titles
Data scientists do everything
Data analysts specifically responsible for data analysis
related to business decisions
Current trend of job titles
Train ML models MLOps and infrastructure work
● Data Scientist ● Machine Learning Engineer
● Machine Learning Engineer ● MLOps Engineer
● Applied scientist ● Infrastructure Software Engineer,
ML
Data analysis, A/B testing
Data Engineering
● Data analyst
● (Product) Data Scientist ● Data Engineer
● Data Scientist in a startup 😛
Week 1, Part 3
The end-to-end
ML interview process
ML interview process
● Application, resume screening
● Recruiter call (behavioral)
● Technical Interviews
● ML theory
● Programming
● Case study (how would you build…)
● Behavioral interviews, cont.
ML interview process (detail)
Source: Machine Learning
Interviews by Susan Shu
Chang (O’Reilly)
Week 1, Part 4
Identify gaps between your current skills
and your target ML role
ML skills matrix
Optional self-assessment survey: Machine Learning
Interviews, Chapter 2
How do you know there’s a gap? (1/2)
● Skills matrix gaps
● Didn’t pass interviews
○ Was there feedback? (Companies don’t give feedback much these days)
○ What’s your self assessment?
○ Even the job titles and descriptions themselves can be telling
How do you know there’s a gap? (2/2)
● Skill matrix gaps
● Didn’t pass interviews
That’s where the ML roles overview helps:
● You could be applying for something that’s the same title, but not a
match for your experience!
● Not your fault, but now you know!
● This training will help you identify and target well in interviews.
How to close the gap between your skills and ML role
Example goals:
● Take ~3 courses
● Build a side project yourself
● Bootcamps… yay or nay?
What’s your curriculum?
● Spend some time to write down:
● 3 Steps to take in the next 6 months
● In 3 months, I will complete…
● In 4 months, I will apply to 5 jobs a day…
● In 5 months, I will reach out to N hiring managers a week…
Week 1, Part 5
Optimize your resume
for target ML roles
Previous trend of job titles
Data scientists do everything
Data analysts specifically responsible for data analysis
related to business decisions
Current trend of job titles
Train ML models MLOps and infrastructure work
● Data Scientist ● Machine Learning Engineer
● Machine Learning Engineer ● MLOps Engineer
● Applied scientist ● Infrastructure Software Engineer,
ML
Data analysis, A/B testing
Data Engineering
● Data analyst
● (Product) Data Scientist ● Data Engineer
Tip: See ML lifecycle graph ● Data Scientist in a startup 😛
Resume sections
Core sections: Optional sections:
● Experience ● Skills summary
● Education ● Volunteering
● Interests
● Other sections
Note regional differences for resumes, such as profile pictures
Example 1
Example 2
Example 3
How to target your resume
● Write down all bullet points for your past experiences (main list)
● For each job posting, take a look at the core ML skills
● Pick a few points from the main list (copy paste) to your resume for
that role
Common additional requirements
● git, version control
● Cross team collaboration
● Writing tests
● (Senior) Ability to break down tasks into smaller ones
Commonly mentioned on JDs, but easy to overlook.
Resume example
(Screenshare)
Week 1
Hands on exercise
Activity: Write down your skills!
1: Beginner
2: Experienced
3: Very experienced
Skill Self identified skill level Skill Self identified skill level
Programming tools Data wrangling
Data visualization and Machine learning
communication (theory, algorithms)
Data intuition Software engineering
Statistics Multivariable calculus
and linear algebra
“Homework”
● Look at the skills you’ve filled out this session
● Identify the gaps
● Create a 3 month / 6 month plan
Take a picture of this slide and the previous!
Next week, we’ll go through…
Overview: Week 2 of Machine Learning Interviews
● Understand the technical fundamentals of machine learning
interviews, such as machine learning theory, statistics, and coding
assessments. (Week 2)
● Gain skills for acing behavioral interview questions, and how to shape
your machine learning career narrative. (Week 2)
Q&A
Learn more at:
ML Interviews site:
Susanshu.substack.com
Get the Machine Learning Interviews book:
https://amzn.to/4aOjO26
LinkedIn:
linkedin.com/in/susan-shu-chang/
