Hello friends!
You’ve now learned quite a bit about what AI is, how it works and such. Next up to learn is how it can work for YOU and your job and life. We believe that the best way to do that is to look at people doing it now!
Let’s get into the 5 kinds of roles people do, who build AI products. These might be fields you’re already in OR they might be roles you are interested in moving into. For each, we’ll also share a link to a recent job posting for that kind of role. There are LOTS of companies hiring for these kinds of roles so if you are interested, now is a good time to look at the requirements. Plus it’s a good time to start getting skilled up in the areas you’re not super-confident in.
Keep in mind that this is a NEW field so getting into it today is a very good plan to future-prep your job.
The roles we will be talking about are:
- Machine Learning Expert
- Problem Space Expert
- Data Expert
- Engineer/Maker/Builder
- Validator/Red-Teamer
1. Machine Learning Expert- this is traditionally the person we think of when we think of an “AI person.” These are people who likely have been doing machine learning work for decades in traditional AI. Currently, they might be doing roles where they train AI models, fine-tune AI models, and document what AI models do. They likely studied calculus, stats, machine learning and such.
If you are curious about this role, try doing a free course, like the Intro to Machine Learning course on DeepLearning.ai
If you do this kind of role now, please document what your AI model is good at with prompt examples and answers. Otherwise the rest of us are playing a giant guessing game: is your model good at math? Code? Language? Images? Or what?!
Example job: Open AI is hiring lots of machine learning folks all over the world! https://openai.com/careers/machine-learning-engineer-applied-ai/
2. Experts in the Problem Space – this can be lots of job titles like Product Manager, Business Development, Business leader, Sales, etc. Basically you deeply understand your target customers’ pain points and are trying to figure out if AI is a good solution or not. This is one of the MOST important jobs in the AI-verse because your job will be to define IF AI is a good solution for the problem and if so, WHEN to use it. In this role, you likely define the problem, you define the solution and then you define ideal and acceptable solutions.
If you are curious about this kind of role, you can do this as a side-hustle to your job. Identify a customer group you understand very well (likely you are part of this group) and figure out what in an ongoing pain-point for them. Once you do this, identify a potential solution. Then figure out if Generative AI is a solution to the problem. Document this whole journey.
If you do this kind of role now, please write some things down about the specific business process you decided to use AI for, what kinds of metrics you want to move (reduce cost to do X by Y) and if AI was the right solution or not. Everyone is looking to you for business cases for AI.
Example job: Headspace (the mental health app) is hiring a product manager https://www.linkedin.com/jobs/view/4104384278
3. Data Expert – This can be job titles like Data Engineer, Data Science, Data Analyst. In some combo of these, your job will be to get the data in order and analyzed so it can be used to “AI over”. Good data is the biggest pre-requisite to any sort of useful AI. If the data is out-of-date or wrong, the AI results will also be wrong. Your role might include identifying the right data sources to AI over, identifying issues in the data such as gaps, insecure data, out-of-date data, organizing the data to be AI-able and keeping this up to date. Another new activity is setting up certain restrictions around your data estate so certain data sources can be left out of the AI party.
If you are curious about this role, there are tons of free data science courses for you to take. Learn who will survive the Titanic (IYKYK!)
If you do this kind of work, please write down your best practices around Data Security because that is the #1 question we get!
Example job: Have a look at this job at Anthropic where they are hiring for a data expert to focus on trust and safety https://jobs.therundown.ai/jobs/100230550-data-science-and-analytics-trust-and-safety
4. Solution Builder – You might be a dev, a maker, a low-coder, etc. You build things for a living, working closely with the business expert and the data person. In the AI-verse your job will STILL be mostly following good engineering principles with the addition of figuring out how best to call the AI model of choice for specific tasks. Which model to use for what? What temperature and other settings to use? How to make this rigmarole performant? In this role you will co-design the solution, you will research and prototype potential solutions then you will get user feedback and refine your solution.
If you are curious about this kind of role, grab a low-code AI tool like Microsoft’s Copilot Studio and go to town and build a chatbot for some real problem you have.
If you are doing this kind of role, please try all the things: metaprompting, RAG, plug-ins, connectors, etc THEN hassle the machine learning people to fine-tune models!
Example job: A new job is at Lakera AI for a senior software engineering for AI. You will find LOTS of AI engineer jobs out there right now. https://jobs.therundown.ai/jobs/100306627-senior-software-engineer-zurich
5. AI Validator/Red-teamer – Okay new job time! Well, kind of. Red-teaming is a term that the AI-verse borrowed from the Security industry. AI Validators/Red-teamers are new roles where people look at the AI products that are being created and they ensure the solutions don’t have strange biases (this is what we see in the news all the time), the product deploys to the right people (not all of us need a paralegal Copilot) and they teach people to use the AI product (this is the hardest part of this whole thing)
If you are curious about this kind of job OR doing this kind of job, please create a prompt library and an AI Skilling program for your space!
Example: This contract position to do AI red-teaming is pretty interesting! https://www.linkedin.com/jobs/view/4099115626/
Okay, those are the jobs that exist today in people who want to be a part of building AI products and systems. Next week, we will look at AI Power Users — people who use AI to do specific job functions.
💗 Dona and Jeremiah
Leave a Reply