Gone are the days I would proudly tout the advantages of working in a product company—as I used to at the drop of a hat, reveling in the pride of having a point of view. After all, POV ^ Skillset = Confidence, isn’t it? However, having POVs are not sufficient on their own. They need to be backed by knowledge/skillsets to get you the required confidence. Also, funnily enough, having one POV doesn’t get you anywhere. Having multiple POVs is important, and thus it is good practice to question your POV and gather more of them as you grow. With high growth comes the need to support scale and efficiency. Some examples that immediately come to mind are 1) Scaling a proof-of-concept to a full-fledged application; 2) Scaling a solution to work for various regions, products, and so on; and 3) Scaling internal practices to make growth seamless.
The AI & ML world (I know this is a broad term) faces a complex set of problem statements that urgently require codification and automation. If we visualize analytics problems by breaking them down into various steps, each step has its own opportunities for codification/automation. Some examples of steps are DevOps (setting up the required infrastructure to support storage and computation), data engineering (ingestion, pipelines, schema), or data science (transformation, EDA, modelling). Not all of them fit nicely in the mold, and for some processes, codification is something we should certainly aspire for.
Combine this with consulting and you are served with the problem statement of customizing solutions based on customer outcomes. The ability to support this type of customization requires us to be flexible and open to pivoting on ideas, but it should also stop us from force-fitting solutions into spaces they will not work in. For example, a solution that requires manual intervention during modelling will not work for an organization if you have hundreds or thousands of models to build.
All of the above factors are amazing directional pillars. They give us well-qualified real-life problem statements with dollars attached to outcomes. Also, we have access to stakeholders from Day 1, and that is who we are truly working for here. We build solutions to transform our partners’ analytics capabilities. For example, if I see that Prophet as a modelling technique is being used in multiple forecasting problems in the organization, we are responsible for taking a crack at standardizing the usage of Prophet across the organization and maybe even set up a reusable component there.
We have an amazing pool of mega-talented data scientists, data engineers, and business analysts who have solved various real problems at a global scale since analytics has emerged as an industry. Core industry knowledge and empathy for our clients help us keep our solutioning approaches honest. Deep technical knowledge on data engineering has helped us create reusable objects that will help us scale and customize. Further, deep design and math knowledge in data science have helped us ensure that these solutions will reach production. Finally, the nimbleness we have ingrained in terms of experimenting-failing-learning holds us in good stead and is an amazing asset—one I greatly treasure at MathCo.
Usually, all envisioned solutions at different levels of maturity need feedback. Whether it is feedback on the approach, modelling technique, scaling, implementation, or even UI/UX, it is a real pleasure to interact with internal industry experts and company leadership, who bring in a vast repertoire of client empathy and real-world problem-solving experience. This helps us create a collaborative and iterative approach towards creating solutions while keeping the end outcome in mind. And since all the work we do is grounded in real problems with real stakeholders, there are rarely any throw-away features. Usually, I am used to approaching this information collection process through external surveys, workshops, and feedback sessions. However, approaching these sessions backed with the arsenal of experience is very essential and helps us make clean, good choices.
We are in a fast-changing world of data and data problems. We need new ways to look at solving old problems or sometimes need to change the problem definition altogether. Our product engineering team is front and center in this endeavor and helps us create solutions that are ground-breaking and truly moonshot. This feeds into both marketing and delivery seamlessly, thus making sure that we keep pace with the changes in fields of technology and data science. Whether it is Alexa-enabled on-demand analytics chatbots or VR-enabled customer heat maps, we are able to effectively consider these solutions for existing problems.
The decision to move to a high-growth AI & ML consulting firm was a no-brainer when I saw the facts spelled out loud and clear. If you feel the same, we look forward to adding your experiences to our ever-growing team of data engineers and data scientists.