Minimum Viable Skillsets (MVS) required to build AI & ML products
In the recent decade the number of products utilizing AI/ML to solve problems have grown exponentially. Every product produces data and organizations of all sizes are investing into smarter ways of utilizing this data. If you are a new CEO or product owner developing a machine learning or AI based product, one of the most confusing things is to figure out the core skills you need to get a solid first version of your product ready. These are my recommendations for what you need as the minimum viable skillsets (MVS) in your team:
Applied Data Scientists
(Yes, there is more than one kind of data scientists)
Who they are: Have experience in applying data science to a real problem whether in their past job(s) or academic degree/boot-camps etc. Know how to code but most importantly know how to convert a business problem to a data science problem and never lose sight of the former.
What they care about: These data scientists are focused on bringing solutions to solving real world problems and hunker down on the user. They design and build solutions keeping in mind the constraints of the product workflow scenarios.
How to expand: Bring in research data scientists to fuel the innovation and provide deep expertise on state of the art ML solutions for the kind of business problems the applied data scientists are solving.
ML Engineers
Who they are: Past experience in software engineering or bringing data science workloads to life in production. Mainly they understand the pains of running and maintaining successful software code in production.
What they care about: Ensuring that solutions designed by data scientists successfully see the light of the day and eventually exhibit all characteristics of a good working software (scalable, repeatable, extensible, you know the drill).
How to expand: Bring in software engineers to build utilities from scratch or on top of existing ML tools that enable faster development and iterations for ML engineers on your product.
Data Engineers
Who they are: They live and breath data and dream in querying languages (okay not necessarily). They have experience working with datasets of characteristics similar to your product. Bonus points if they have worked with datasets from the same industry as that of your product.
What they care about: Giving makeovers to datasets such that they become worthy of use by everyone else. Also how to do the same at scale in production.
How to expand: Bring in data scientists to work with data engineers such that more and more of the data preparation gets baked into your data pipelines giving all the more time and freedom to data science teams to focus on the rest.
UX Designers
Who they are: People who design experiences of tech products with or without a user interface. They can effectively instill the user experience in every phase of development with their technical peers and ensure that the glamour of complex technology doesn’t take away the usefulness of the product.
What they care about: Simplicity, usefulness, precision — Threading in a seamless user experience in every bit of the product.
How to expand: Bring in user experience researchers and interface designers to truly determine both explicit and latent needs of your users.
Domain Experts
These are important if you are solving a specialized problem
Who they are: The actual user or someone who knows the user so well, you can’t tell the difference. They can problem solve with other team members and brainstorm, guide, validate, test your AI solutions.
What they care about: Providing specific domain insights and expertise to help build solutions that will land well with the users.
How to expand: Bring in more user experts. The bigger representation of the actual user in building the product, the better your product will cater to the actual user base.
Product Managers
Who they are: Have experience in building technology products with bonus points for AI/ML based products. Background can be technical, business, domain user etc., what is most important is they understand the depth and breadth required for building a successful product and are willing to go the lengths to meet that
What they care about: Solving the user problem. Period.
How to expand: Bring in product managers to focus on growth, marketing, release and customer support of the product amongst other specific areas. This way your product has all the wings it needs to fly!
The focus here was on the minimum viable skillsets (MVS) needed to develop your AI/ML product. Depending on your industry and business needs, you may need additional or different skillsets to develop and ship your product. Another important thing to note is each skillset does not equal a separate individual per say. You might find people who possess multiple skillsets together.
I propose you define what your MVS looks like and adapt as you go. Best of luck!
