Cracking the AI/ML Product Manager Interview

Original article via Product School Blog by Adrian Gonzalez Sanchez https://productschool.com/blog/product-management-2/cracking-ai-ml-product-manager-interview/ (October 10, 2021)

This article is an overview of the current state of the AI/ML product management roles around the world, along with a series of recommendations to succeed in cracking the AI/ML PM job interviews. It includes personal notes based on the author’s industrial and academic experience.

Before we start, here is the industry context, because you need to be aware of what is happening out there in order to prepare for your future job interviews: The increasing level of AI maturity of adopter organizations around the world is both a) consequence of the granularity of tasks within well organized AI teams, b) a motivation to continue hiring to create talented multidisciplinary teams.

Part of this granularity relies on profiles such as ML engineers, AI ethicists, and of course our beloved AI/ML Product Managers. This was not the case some years ago. Most of the hiring efforts were fully dedicated to data science and engineering folks, but good news for you… there is life beyond the typical data trilogy (data science, engineering, and analysis), and the AI/ML product manager is now part of the family, and a standard for a lot of companies.

This is actually an amazing source of opportunities for companies and candidates, but also brings a crazy mix of unstandardized interviewing and hiring practices. A sort of wild west where interviews for similar roles in two different companies will look and feel totally different, depending on the organizations, the hiring managers, and other involved stakeholders.

Now, from an “AI/ML PM” candidate point of view, how to navigate this uncertainty and make the most of the situation? Can we become experts and crack these kinds of interviews regardless of which company we are applying for? Let’s see…

What is actually an AI/ML Product Manager?

As usual, the best answer is “it depends”. For good or for bad, there is a panoply of interpretations of what AI/ML product managers and their skills look like. Based on this author’s experience, there are several factors (besides those usually associated to normal PM roles) that will influence the companies’ perception of what an AI/ML PM and its hiring pricing should be:

• Level of maturity of the organization in terms of AI/ML development — Try to read the current stage of the AI/ML efforts within the organization. It is important to understand where they are, in order to adapt your value proposition and to better sell your experience. For example, I have met a lot of companies who were looking for an AI/ML PM, but in reality they wanted a coach for the rest of the team. This is fine, but it requires good research and analysis skills, before and during the interviews.
• Executive support and involvement — Executives and even advisors can play a crucial role. AI product managers are usually their connection with their technical teams and they are looking for a mix of doers and translators. This means, getting things done and having the ability to discuss progress, ideas, roadblocks, and all kinds of relevant topics with them. In this case, communication skills are gold.
• Role of the HR department within the organization — This will guide not only the type of interview process, but also the potential benefits you may get if they hire you. I have seen a lot of companies looking for AI/ML PM impressive backgrounds (they are usually the result of several existing job descriptions, what I called an “AI unicorn”), but offering standard or even low salaries. This is the case for companies where levels are granular and fixed by human resources.
• Existing technical people within the team — Keep in mind that you can join a company before even there is a proper AI team. The AI/ML product manager can arrive after one or two data scientists, but the pattern I see the most is the PM being the 3rd or 4th, right after scientists, data engineers, and — depending on the case — some data analyst).
• The technology environment — The kind of IT infrastructure the company has will also impact the interview process. It is not the same to implement AI/ML models with the cloud, or doing everything on premises. I see an interesting trend in which companies are expecting AI/ML product managers to understand and even master some of the public cloud technologies (e.g., GCP, Azure, AWS).

Obviously, other general questions such as the level of the existing PM practice, the industry and its competitive context, and the seniority of the candidates are also relevant and will feed the discussions during the interview process.

Interview scenarios based on the type of company

Based on the previously listed factors, we can classify the interview process according to a series of potential scenarios, all of them based on real experiences with hiring companies and clients:

• Pragmatic interviews — This is one of the best case scenarios. It usually happens when the companies really know what they are looking for, which obviously requires a high level of internal AI/ML and product management maturity. Their interview process is usually standardized with a series of predefined interviews. I have seen that with big tech companies who are already implementing AI/ML and bringing new resources to grow the team.
• 360-degree interviews — This scenario is not necessarily bad, but it requires a lot of preparation and multidisciplinary knowledge. A lot of companies are adopting this while looking for their perfect AI/ML product managers. This is usually the case for smaller adopter companies which are trying to find the right person for the role, usually when they are hiring their first AI/ML PM. For them to be totally convinced, they usually prepare a series of interviews with all kinds of stakeholders, from top executives to very technical folks.
• Niche interviews — There are different situations that will depend on the context and expectations of the firms. I have seen companies looking for very specific B2B or B2C experience, where the AI/ML knowledge requirements were almost testimonial. Others, as it usually happens for PM roles, focus a lot on the industry experience — e.g. They can prioritize someone with experience in e-commerce or fintech, and accept that the AI/ML knowledge will come later. I have seen this kind of scenario with AI consulting firms too.
• Technical exchanges — Relevant for most of the AI/ML PM roles, it is very usual to be interviewed by data scientists, ML and data engineers, or others, just to validate that the candidate understands the fundamentals and can speak the same lingo. Based on my experience, this can go as technical as the candidate wants, and it is a good way to start developing trust with the technical team, even before joining the company.
• HR-oriented process — It really depends on the role of HR during the interview process, and their actual understanding of the required soft and hard skills. Based on my experience, you can face situations where it will be difficult to connect your background with the keywords and specific ideas they may have in mind. Navigating this is almost an art and it requires a great ability to explain your experience in simple words. AI/ML is unknown for most people, and HR is definitively not an exception.

Recommendations for interview preparation

In summary, this new wave of AI/ML product management opportunities brings challenges too, but there are ways to crack the interview and get the job. Here are a series of recommendations based on my past successes, and obviously on the painful failures:

• Build your unique value proposition mix — It really depends on your professional and academic background, but I usually recommend to connect at least to aspects that will make you special and better than other candidates. In this case, it could be industry-related experience, previous AI/ML experience in similar projects, IT architecture skills to help design AI/ML solutions, teaching experience that you may leverage later with your colleagues and clients… everything counts, and you will have to reflect on that and build your own mix.
• Bring your pre-interview research to the next level — This is the case for any kind of job interview, but it becomes hyper-relevant for AI/ML PM interviews. There are just too many moving parts, and everything is about reducing the risk of failure. Proper research may include previous papers, articles, technical documentation, social media activity from existing team members, recorder executive talks where the C-suite shares pieces of the corporate strategy and priorities. One of my recommendations to my university students is to force themselves to pre-fill both a Business Model Canvas and SWOT analysis so they can fully understand the internal and external context of the company. This rich analysis will generate better discussions and give a sense of seniority to the hiring managers.
• Be ready for case-based interviews — I have seen this a lot with hiring companies and clients, where they quickly present a situation or opportunity, and they ask candidates to elaborate and plan and to present it to a panel. This is always relevant for product management positions, but AI/ML is a specially good area. Why? It gives you a perfect opportunity to showcase your knowledge, and it can be a huge differentiator if you are better than other candidates. That being said, get ready for all kinds of cases, from single paragraphs with open ended questions, to fancy cases asking you to create a roadmap for a specific AI/ML solution.
• Be a proactive teacher — The best interviews are those where you can bring your knowledge and explain different topics to a diverse audience (executives, HR, technical teams). This one is a winner, and it will require knowledge, experience, and excellent communication skills.
• Develop multi-level communication approaches — Related to the previous point. An AI/ML product manager will need to have different business and technical discussions with all kinds of internal and external stakeholders. This requires a lot of preparation and versatility, but once you master it, it will be a powerful asset during the interview processes.
• Go as technical as you can — Even if the technical skill requirements depend on the company, no one will complain if you show a solid technical background. It is true that I have lived situations where my interviewer will think I’m “too technical” for a product manager, but those are exceptions. If you learn and practice something about AI models, cloud computing or MLOps, why not to leverage it during the interviews?
• Showcase the portfolio of past projects — The AI/ML portfolio is the king. Similar to artists, companies will hire AI/ML talent and pay higher salaries to people with previous AI/ML project experience. In this case, it could be a combination of professional and academic projects, decorated with some nice certifications and badges.

• Consolidate your AI/ML story — Again, this is not only specific to AI/ML product managers, but it is a general requirement to succeed, not only during the interviews but also once you join the company. It is important to be aware of your own valuable experience, and to showcase in the form of relevant examples or advice during your team meetings.
• Leverage new AI/ML frameworks and canvases — It is not easy to organize processes and knowledge for new and complex topics, but there are new tools out there that will help explain AI/ML projects and use cases. Some examples: AI Canvas, ML Canvas (more technical), Open Ethics Canvas, Big Data Maturity Index.
• Be honest — The last but definitely not the least. It is the 101 of any interview process and something I have always applied during my interviews, but it makes even more sense for AI/ML (PM or not) positions. We are continuously learning, facing unknowns, using new tools, etc. There is no need to pretend we know all because no one does. It is impossible to be a product manager, data scientist, designer, architect, and everything at the same time. But it is good to explain our strengths and limitations, and to show a plan to keep growing and learning. No need to lie or to exaggerate.