What types of roles are available in AI
Types of data you'll work with. Responsibilities of each role. Job titles within each type of role.
Hey reader, welcome to the 💥 free edition 💥 of my weekly newsletter. I write about building startups and careers in AI. You can learn more about me here. Feel free to send me your questions and I’m happy to offer my thoughts. Subscribe to this newsletter to receive it in your inbox every week.
AI job market is on fire right now. Just look at the number of companies hiring for AI roles! If you want to enter the field from another field or you're just starting out, you need to plan your career and figure out what you want it to look like. What are the different types of roles in AI? What type of data do you want to work with and what role is right for you?
I've listed 6 types of roles based on the type of data/environment you'll work in:
Numbers
Images and Video
Audio
Text
Hardware
Ops and Architecture
Here’s what each type of role entails:
1. Numbers
You'll work with numerical data coming from various sources. Professionals who can do this are in high demand and many companies hire for this role such as Google, Amazon, Microsoft, Tesla, Apple, and more.
Types of data: sensor data, user data on web apps, user data on mobile apps, population data Any company that collects numerical data needs people who can make sense of it. Your job is to build models that can represent this data and extract meaningful insights from it.
Roles: Data Analyst, Data Scientist
2. Images and Video
Applying ML to images and video comes under Computer Vision. You'll work with image and video data. The goal is to extract information from these images and videos so that you can take further action.
Types of data: image search (Google), streaming (Netflix), image sharing (Instagram), video sharing (YouTube), identifying cars on the road (Waymo), face recognition (Apple iPhone).You job is to build models using image and video data. These models are used to detect and recognize content in images/video. We experience this technology every single day in our lives. It has become ubiquitous.
Roles: Computer Vision Engineer, Video Processing Engineer
3. Audio
You'll work with audio data. The goal is to extract information from audio data so that you can take further action based on the outcome.
Types of data: mobile phones (Apple Siri), connected devices (Amazon Alexa), recognizing music (Shazam), recommending music (Spotify), speakers (Sonos)You job is to build models using audio data. These models are used to detect and recognize content in audio. We experience this technology every single day in our lives. It has become ubiquitous.
Roles: Audio ML Engineer, Speech Recognition Engineer
4. Text
Applying ML to text data comes under Natural Language Processing (NLP). You'll work with text data. You have to understand how to parse text data and build models to understand the content.
Type of data: search engines (Google), user-generated content platforms (Facebook, Discord, Reddit), review platforms (Amazon, Yelp), or any place where large amount of text is being generated by the users.You have to build models to understand the context of sentences. This will enable you to converse with humans.
Roles: NLP Engineer, NLP Researcher
5. Hardware
You'll work with hardware products like robots and autonomous machines. Or you'll work on hardware products like GPUs and chipsets to accelerate AI algorithms on mobile, edge, and cloud.
Types of data: robots (Tesla), GPUs (Nvidia), image processors on mobile (Apple)Companies like Apple, Tesla, Nvidia, Intel, Google, and Amazon are hiring all the time for this role.
Roles: Robotics Engineer, AI Hardware Engineer
5. Ops and Architecture
You'll work on building the operations and architecture for AI products. This is mostly centered on efficient storage, movement, and retrieval of large amounts of data.
Type of data: It’s applicable to all types of data For example, what's the right architecture for a company that has to stream video to millions of users? Or what's the right architecture for a company that has to answer large number of text queries every second? It's a fluid field and requires a lot of experience.
The requirements keep changing as the company scales up, so the architecture needs to keep up too. Inefficient data/AI architectures can cost millions of dollars to companies on a daily basis. Every modern tech company hires for this role such as Netflix, Google, Apple, Amazon, Facebook, Nvidia, and more.
Roles: Data Engineer, AI Architect, Data Infrastructure Manager
Where to go from here?
You need to remember the magic of compound interest when you’re building a career. Use it to your advantage. Don’t do a bunch of unrelated things hoping to land somewhere. Decide what you want to do early on and build upon that. This will enable you to build expertise and build a career you’ll be proud of.