Practice and guide
Collection of problem solving and system design questions. General guidlines from interviewers perspective.
Machine Learning Roles
Although not a comprehensive list, ML roles can be split into three roles that complement each other in a typical project lifecycle. These roles are:
Applied Scientists: Applied Scientists leverage data to identify and tackle opportunity areas, employing statistical inference and machine learning skills in exploratory as well as production settings to develop new products, optimize business metrics, and perform causal inference.
Machine Learning Engineers: Machine Learning Engineers work on engineering ML systems such as Yelp’s real-time ad targeting and search ranking services or batch-based content mining applications. This involves an understanding of software engineering and system design, data ETL and data stores, and machine learning.
ML Platform Engineers: ML Platform Engineers build the infrastructure and tooling to support our Machine Learning Engineers and Applied Scientists.
Resume screening
What recruiters/hiring managers look in your CV?
Google Resume Tips and Advice
Top: Programming languages, github profile
For experienced industry professsional list:
- First section: Experience should be the bulk of the CV.
- Second Section or within First Section: Projects
- Leadership and awards section
- Final Section: Education
List everything in reverse CChronological order with your most recent experience first. Use action words: created, dsigned, debugged, negotiated,developed, managed.
Framework: Accomplished [X] as by measured by [Y] by doing [Z]
Increased server query response by 15\% by restructuring API.
Grow revenue from small and medium business clients by 1- \% by mapping new software features as solutions to their business goals.
Your CV
Example LinkdIn profiles of AS/MLE people with no Phd:
-
https://www.linkedin.com/in/shanchaoli/ A lot of projects. His DS work is closely related to productionizing ML pipelines. Also takes part in business definition of the problems using ML.
-
https://www.linkedin.com/in/tianwng/ - Publications...
-
https://www.linkedin.com/in/makeshsreedhar/ - no phd but has showed doing some research internships
MLE:
- https://www.linkedin.com/in/alex-lo-19528414a/ (DS/DA)
- https://www.linkedin.com/in/aman3014/ (CS, Computer Vision)
- https://www.linkedin.com/in/rayan-hatout/ (AI/ML)
- https://www.linkedin.com/in/thakarar/ (DS, SE)
- https://www.linkedin.com/in/dtomek/ (Applied Maths + SE)
- https://www.linkedin.com/in/yunru-l-49157784/(Maths + Engineering)
- https://www.linkedin.com/in/simone-merello-bb0aa6107/ (CS)
Looks like for Meta to become a Research/Applied Scientist you need a phd. Out of 30 profiles I found only 3 without a Phd.
For MLE, looks like you do not need a Phd. Most of the MLE people have Bachelor or MS. However, they all show proof that they know how to code.
No fluffy things!
Carefully select which projects to share from your Github (remove snake and ping pong games).
General guidelines
How to grade Problem Solving Interviews
Problem Solving Interview Format
Applied Scientist
Problems
ML/AS
Full stack, App Backed, Data Backend
Distributed Systems
SRE
ML System Design
AS often don't have engineering backgrounds, so we don't expect them to have depth in infrastructure or engineering oriented areas. Eg, MLEs should know about caching and risks of different online vs offline feature code paths; but that's not required for AS.
AS Ads Experiements Metrics System design
Programming principles
System Design
The focus of the System Design interview is to evaluate whether the candidate can take a somewhat high-level problem, understand the technical requirements and propose a concrete system that solves the problem. While it’s nice if the candidate can solve all parts of the problem, it’s more important for the interviewer to focus on asking good questions and guiding the candidate in order to evaluate the candidate wrt all the criteria that would make them a good “system designer”.
All interviews follow the same format: a short introduction, followed by a significant amount discussion, brainstorming, and when relevant some "white-boarding".
[1-2m] Introduction Introduce yourself and your team.
How does the work you do impact Yelp?
Mention a few things that make you excited about working at Yelp.
Go over the format of the interview:
Few minutes discussing the candidate’s interests Bulk of the time on a technical problem involving coding and design Five minutes at the end for questions about Yelp
[3-5m] Interests Learn the communication style and build a rapport with the candidate, so that you can better understand them later on. This also helps make them more at ease with the interview format. Note that information from this section isn’t incorporated into your rating, but you can call out any concerns you observe here in your written feedback.
Explain what this interview is about and that it will include little to no coding but will be more of a discussion and possibly some whiteboarding.
Learn a little of their history from them, past experience, and projects. What types of systems have they worked with?
[35m] System Design Use one of the 2 questions from the pool of questions in your track that you have familiarized yourself with. This is rare but if it is not the candidate's first SD interview, make sure that the question you chose hasn’t already been used in the previous interview.
Each system design question has all the details about how to pace the interview, how to nudge and what to focus on. Below are the system design tracks linking to their corresponding SD questions:
[5m] Questions for the interviewer Allow a few minutes at the end of the interview to allow candidates to ask you questions.
Fullstack, Data backed questions for SD are the same as Backend questions.
OTC (Ownership, Tenacity and Curosity)
Behaviour (play well with others)
Additional Resources
Hacckerrank questions See folders:
- questions
- structured questions (System Design)
Misc
I made the quiz based on FAANG interview exp here: https://github.com/khangich/machine-learning-interview
motivational videos
Activation energy - try set alarm early
its your job to do the crap you need to do to become what you want
you will never feel it... you have to FORCE yourself
Your brain has two states autopilot and emergency break. Whenever you do something different from your routine your brain would go EMERGENCY BREAK. You need to force yourself to overcome it.
Its the routine is what is killing you. If you feel sad or dissatisfied it is a SIGNAL, that some of your basic need is not met. You need to FORCE yourself to do something different. Get out. Get ouside of your comfort zone.
When you have an impulse you need to ACT. 5 secods rule, if you wait more than 5 sec you are unlikely to do anything.
I know there are a lot of people out there affected by layoffs and I hope this post gives you some inspiration that the job market isnt as scary as people make it seem. I ended up with 2 offers:
Netflix Senior SWE: 485k base salary (a bit low compared to what I’m seeing online but I’m happy)
Tiktok: 250k base 25% avg bonus 100k sign on bonus split over 2 years and 400k in RSU (paper money)
I still have an on-site left with Jump Trading which if it goes well I can assume tc above Netflix.
Edit: Resources for those who need them:
Referral site: Refer.me
Resume ATS checker: Skillsyncer.com
Resume generator: Https://resume.lol
Mock interviews: Interview.io
I have 6 YOE and my previous TC was 280k so I very excited for what the future holds 😊 lmk if there are any questions I can answer as well
Backlinks