Elevate your technology and enterprise data strategy to Transform 2021.
How do you start to build an AI team? Well, hire unicorns who can understand the business problem, can translate it into ‘good’ AI building blocks, and can take care of the implementation and deployment of production. It sounds easy! Except that sightings of such unicorns are extremely rare. Even if you find a unicorn, chances are you can’t afford it!
In my experience at the helm of Data + AI products and platforms over the past two decades, a more effective strategy is to focus on recruiting strong performers who cumulatively support seven people with specific skills in the field. ‘team.
The 7 Skill Characters of a Unicorn AI Team
Persona data set interpreter
The cornerstone of an AI project is data. Finding the right datasets, preparing the data, and ensuring consistently high quality is a key skill. There is a lot of tribal knowledge about datasets, so you need someone who can specialize in tracking the meaning of data attributes and the origin of different datasets. A related challenge with data is tackling multiple definitions within the organization for business metrics. In one of my projects, we were dealing with eight definitions of “new monthly customers” in sales, finance and marketing. A good place to start for this skill figure is a traditional data warehouse engineer who has strong data modeling skills and an inherent curiosity in correlating the meaning of data attributes with application and business operations.
Pipeline builder personality
Getting data from multiple sources to AI models requires data pipelines. In the pipeline, data is cleansed, prepared, transformed, and converted into ML functionality. These data pipelines (called Extract-Transform-Load or ETL in traditional data warehousing) can get quite complicated. Organizations typically have pipeline jungles with thousands of pipelines built using heterogeneous big data technologies such as Spark, Hive, and Presto. The Pipeline Builder persona focuses on building and running large-scale pipelines with the right robustness and performance. The best place to find this character are data engineers with years of experience developing batch and real-time event pipelines.
Persona AI full-stack
AI is inherently iterative from design, training, deployment and retraining. Building ML models requires hundreds of experiments for different permutations of code, features, datasets, and model configurations. This character is a combination of knowledge in the field of AI and strong skills in building systems. They specialize in existing AI platforms, such as Tensorflow, Pytorche, or cloud-based solutions such as AWS, Google, and Azure blue. With the democratization of these AI platforms and the generalization of online courses, this character is no longer a rarity. In my experience, a solid background in software engineering combined with their curiosity to gain a mastery of AI is an extremely effective combination. When hiring for this character, it’s easy to meet geniuses who like to fly solo instead of being a team player – be on the lookout and take them out early.
Personality of AI algorithms
Most AI projects rarely need to start from scratch or implement new algorithms. The role of this persona is to guide the team through the search space for AI algorithms and techniques in the context of the problem. They help reduce dead ends with heading correction and help balance the accuracy and complexity of the solution. This character is not easy to obtain given the high demand in places focused on AI algorithmic innovations. If you can’t afford to hire someone full-time for this skill, consider bringing in an expert as a consultant or start-up advisor. Another option is to invest in training the entire team by giving them time to learn research advances and internal algorithms.
Data + AI operations personality
Once the AI solution is deployed in production, it should be continuously monitored to ensure that it is functioning properly. Many things can go wrong in production: data pipeline failure, poor data quality, under-provisioned model inference endpoint, drift in the accuracy of model predictions, uncoordinated changes in metric definitions profession, etc. This character focuses on creating the right oversight and automation to ensure seamless operations. Compared to traditional DevOps for software products, Data + AI Ops is considerably complex given the number of moving parts. Google researchers correctly summarized this complexity in the form of the CACE principle: Change Anything Change Everything. A good place to start for finding this character is with experienced DataOps engineers who aspire to learn the Data + AI space.
Hypothesis planner personality
AI projects are full of surprises! The shift from raw data to usable artificial intelligence is not a straight line. You need flexible project planning – adaptation based on evidence or rebuttal of assumptions about datasets, functionality, model accuracy, customer experience. A good place to find this skill figure is in traditional data analysts with experience working on multiple concurrent projects with tight deadlines. They can act as excellent project managers given their instinct for tracking and paralleling assumptions.
Personality of the impact owner
An impact owner is fully aware of the details of how the AI offering will be deployed to drive value. For example, when solving an issue related to improving customer retention using AI, this character will have a full understanding of the journey map associated with acquisition, retention and customer attrition. They will be responsible for defining how the AI solution’s customer attrition predictions will be implemented by the support team specialist to reduce the churn rate. The best place to find this character is within the existing sales team – ideally, an engineer with a strong product instinct and pragmatism. Without this persona, teams end up building what is technically feasible rather than being pragmatic about what is actually required in the end-to-end workflow to drive value.
To sum up, these seven skill characters are a must have for every AI team. The importance of these personas varies depending on the maturity of the data, the type of AI problem, and the skills available with the larger data and application teams. For example, the persona of the data interpreter is much more critical in organizations with data in a large number of small tables compared to those with a small number of large tables. These factors should be taken into account in determining the appropriate seniority and cardinality for each of the skill characters on the AI team. Hope you can now start building your AI team instead of waiting for the unicorns to show up!
Sandeep Uttamchandani Chief Data Officer and VP of Product Engineering at Unravel Data Systems. He is an entrepreneur with over two decades of experience building Data + AI products and author of the book The self-service data roadmap: democratize data and reduce viewing time (O’Reilly, 2020).
VentureBeat’s mission is to be a digital public place for technical decision-makers to learn about transformative technology and conduct transactions. Our site provides essential information on data technologies and strategies to guide you in managing your organizations. We invite you to become a member of our community, to access:
- up-to-date information on the topics that interest you
- our newsletters
- Closed thought leader content and discounted access to our popular events, such as Transform 2021: Learn more
- networking features, and more