The Different Roles You Need to Hire to Build a Successful AI Organization
People matter, timing matters, and org structure matters.
Whether you’re building an AI organization from the ground up at a startup, or you’re looking to establish successful AI projects for the first time at a large enterprise organization, you are probably wondering who to hire for your AI project. Many AI organizations get this wrong. So in this article, I talk through the different roles in AI, what value we should expect them to bring to your organization, and when you should hire them.
Building AI teams isn’t easy. The tech is advancing quickly, and so the roles and skillsets must develop with the industry. This means that what a Data Scientist needed to know to bring value to your organization 10 years ago is not the same as what they need to know today. It also means that the roles and skills are being defined under our feet. And as a result, you’ll see a lot of variance in the roles and responsibilities. This industry just hasn’t reached equilibrium yet in terms of the roles and skill sets for AI. Today, many companies are still trying to figure out who to hire and how to organize and empower their AI teams. So I’m here to help you find your way through it.
Don’t make the mistake of hiring a great Data Scientist and not giving them the right tools or data to deliver value.
Don’t make the mistake of hiring a Data Engineer, and setting them up for failure by expecting them to deliver an end-to-end AI project when that is outside their skillset.
Don’t make the mistake of hiring an experienced researcher when all you really needed was a hacky ML engineer to deliver that lean prototype.
So let’s breakdown the who’s who in AI, what to expect from them, and when to hire them
Individual Contributor Roles
In the tech world, we call the people who are working hands on keyboard “Individual Contributor”s, or ICs. There are a few key IC roles required to deliver successful AI projects: Data Scientists, ML Engineers, and Data Engineers.
Data Scientist
Who are they? A Data Scientist is responsible for delivering the AI or ML model. They clean and analyze the data, and use their understanding of complex modeling approaches and the subject matter to choose the optimal modeling approach for the business problem. They will deliver a working and optimized model.
🔎 How to spot a good one: These Scientists require a strong background in mathematics, statistics, and ML/AI theory. They are typically working with SQL to gather their data, and python and jupyter notebooks to iterate on the modeling approach. They have a proven track record of delivering business value through their projects, and an ability to scope their own projects and translate a business problem into an actionable data science project. They should be great at stakeholder communication.
🕙 When to hire one: There are two scenarios when you’d want to bring in a true Data Scientist.
- Your project does not require a software component. The data scientist is delivering insights and reports using AI/ML. This is typically true on internal BI projects. Your project requires a new, innovative, or complex modeling approach, and you want to bring in a specialist to focus on creating only this piece of the project.
- You expect to be able to publish or patent your approach. Note that you will need to empower this hire by bringing in the right Data and ML Engineers to support them.
CV or NLP Researcher / Scientist
💼 Who are they? These Researchers are similar to the Data Scientists described above, but have chosen to take a specialized path and focus on one complex model type. There are two areas in Data Science today that are complex enough that they could require a specialized researcher.
- Natural Language Processing — A NLP Scientist is focused on working with modeling techniques for written text data.
- Computer Vision — A CV Scientist is focused on working with modeling techniques for image and video data.
🔎 How to spot a good one: The skillset is the same as the Generalist Data Scientist. These Scientists are also required to have a strong background in mathematics, statistics, and ML/AI theory. In addition, they should be experienced with and up to date with the latest modeling approaches and technology in their chosen specialization, and able to contribute to the advancement of these fields.
🕙 When to hire one: Your project requires the use or advancement of cut5ting edge CV or NLP technology.
Machine Learning Engineer
💼 Who are they? These Engineers work closely with Data Scientists to transform the AI or ML models into scalable, production capable software. They are skilled in MLOps, and can rake a Data Scientist’s model out of the jupyter notebook and into the real world where it can be used by the business. Without them, you will often have an excellent model created by a Data Scientist that is adding 0 value because your customers cannot cannot interact with it, or it cannot be automatically run to produce predictions.
ML Engineers are also capable of using existing ML and AI techniques and models. So if you’re in the early phase of a project, or if you project does not require a new or complex modeling approach (many don’t), the ML Engineers can create and deploy the model end to end.
🔎 How to spot a good one: ML Engineers require a strong background in MLOps, software engineering, programming, and a deep understanding of machine learning principles and technologies. Do not hire traditional Software Engineers to act as an ML Engineer. Although I’m sure these engineers would be able to figure it out through trial and error, deploying an AI or ML product requires a deep understanding of MLOps core concepts to understand how to properly operationalize a the model.
🕙 When to hire one: Your project requires a software component. You think you can the project off the ground using known and tried and true AI or ML tools and methods.
Data Engineer
💼 Who are they? There is no AI without a Data Engineer, because AI needs data, and these are the Engineers are the people who make your data usable and accessible to the Scientists and Engineers who knows how to turn the data into business value.
A Data Engineer is a professional who designs, builds, and maintains the infrastructure and architecture that allow for the efficient handling and storage of large data sets. Your data is going to be important for model exploration and research, model creation, and maintaining and improving the model over time. So your Data Engineer is crucial at every step.
🔎 How to spot a good one: Essential skills for a Data Engineer include expertise in database systems, proficiency in programming languages like Python or Java, and a strong understanding of ETL (extract, transform, load) processes, as well as knowledge of cloud platforms and big data technologies like AWS and GCP.
🕙 When to hire one: If your project requires data (it does) 🤷♀️
Leadership Roles
Your team will only be as good as it’s leadership. The leadership team will look different depending on the size of your organization or stage of your company, but in general you will have a leadership team consisting of a Product Manger, Data Science Manager, Engineering Manager, and an Architect. Here are the breakdown of those key leadership roles on an AI team.
AI/ML Product Manager
💼 Who are they? An AI Project Manager owns the roadmap. They specializes in overseeing projects focused on artificial intelligence, ensuring that the objectives, timelines, and resource allocation align with the project’s goals.
Part of your Product Manager’s job is to interface with stakeholders, but DO NOT LET THIS MEAN your ICs are isolated from the use case! Data Scientists cannot effectively build a solution for a problem they are not close to or do not understand. The best Data Scientists will also interface with stakeholders and develop a deep understanding for the product they are building, and customer empathy. The worst Data Scientists will tell they don’t have to do this because this is the Product Manager’s job.
Your Product Manager will keep the project on track as well. This is a terribly crucial role especially if you have those not-so-fantastic Data Scientists or Engineers on the team who only want to focus on the tech and science and are not investing in understanding how their work drives the business or helps your customers. (Although I’d argue these archetypes are simply not the best in the business. If you’re a Data Scientist reading this and you want to be fantastic, follow my tips here.)
🔎 How to spot a good one: The best Product Managers I know are EXPERTS in AI and ML core concepts. Although the Product Managers are not coding hands on, they will be ineffective in their role without a deep understanding of the nuances of an ML project, and the ability to articulate what is possible with this technology today.
🕙 When to hire one: When your project or organization is large enough that it needs it.
Engineering Managers and Data Science Managers
💼 Who are they? These are your typical people managers, but for an AI/ML project, they need to have that background to be able to properly guide and grow their team. These are the folks who will hire, retain, and grow the best talent on your team. They are responsible for coordinating the team’s efforts, setting project goals, and ensuring the alignment of AI strategies with business objectives.
🔎 How to spot a good one: Essential skills include a strong background in AI and machine learning, proven leadership and project management abilities, and excellent communication skills to effectively guide and mentor their team while collaborating with other departments or stakeholders.
This is NOT your best engineer on the team. THAT should be your Architect, which we’ll talk about next. The Data Science or Engineering Manager should be your your best people manager on the team. They should be able to align people around a vision, inspire them, and pave a path for them to succeed.
🕙 When to hire one: When your project or organization is large enough that it needs it.
AI/ML Architect
💼 Who are they? Your Architect is responsible for designing and overseeing the implementation of AI systems and solutions, ensuring they align with business objectives and integrate seamlessly with existing infrastructure. They analyze the requirements of a project, propose suitable AI technologies and frameworks, and guide the development team throughout the project lifecycle.
🔎 How to spot a good one: A great Architect is a great mentor and teacher. Although it is the pinnacle of an IC’s career, this is not just an IC role. They need to be a multiplier. Essential skills to be a great Architect include a deep understanding of AI and machine learning technologies, proficiency in systems architecture and design, and proven track record to develop effective and scalable AI solutions.
🕙 When to hire one: I wanted to mention the Architect because this person is so critical to establishing your AI organization properly. It should be a first hire, and if you can’t afford it, you should position one as an advisor.
Customer Facing Roles
These roles are specifically relevant if you’re a SaaS AI company or product. In that case, there are even more AI capable roles you’ll want to pull in at some point. And these ones are lesser known, so I’m including them here to raise visibility. There is a great career path here for our extroverted Engineers and Data Scientists!
Developer Advocate
💼 Who are they? This is the rockstar job of AI/ML for those extroverted peeps, so if you’re passionate about coding, want to travel and live that rockstar lifestyle, you need to hear about this one.
The Develop Advocate’s job is to teach and inspire others to use your AI product. They spend their time creating easy to tutorials with technical depth, attending and speaking at conferences, creating workshops, etc. They guide your users in AI/ML best practices, and they are creating a community of power users while they do it.
🔎 How to spot a good one: They are deeply technical people. They are social media capable. Similar to influencers, they know the strategy to create content and make it go viral.
🕙 When to hire one: If you need to build a community around your product. You need to educate and inspire your users.
Solutions Engineer
💼 Who are they? They are deeply technical people. This is often a pre or post sales customer facing role. I’m including this one because I think it’s not well known to many new folks in the field, but its an incredible growth role. It will rocket ship your career. 🚀
🔎 How to spot a good one: Essential skills for an AI Solutions Engineer include strong programming abilities, particularly in languages like Python or R, a deep understanding of machine learning algorithms and data processing techniques, and the ability to translate business requirements into technical solutions.
They should be passionate about the product they are working with or the problem it solves. Many of the best Solutions Engineers used to be the customer’s themselves, and have become so frustrated or passionate about the challenges they faced, that they want to help other people overcome them
The Solutions Engineers are great teachers and even better listeners.
🕙 When to hire one: You need to bring your customers custom solutions. You need folks on the front lines working 1–1 with your customers to educate and advise them.
So that is the cast of characters to keep in mind when you’re building our your AI Organization. Remember, you’ll want to start with a small lean team hacky of generalists to prove out the concept and deliver a working prototype that can be iterated on. Once these organizations willstart to grow, and you can layer in true specialists to mature the product.
One thing I’ve learned in 11+ years working in AI and ML is that org structure matters! If you want to learn more about when to hire these folks for these different roles, and how to organize your teams during different stages of growth, check my other article I’ll link here.
I hope this article will help leaders and founders get started with AI at their companies and hire and empower the right people. And if you’re just starting in your AI career, I hope this gave some insight to the different kind of roles you could be hired for, and what might be a good fit for you if you choose a career in AI!
Let’s go build it! 🚀
