From: aidotengineer

Building a GenAI platform requires a dedicated and skilled team [00:00:59]. The approach to hiring for such a team has evolved due to the unique nature of GenAI systems [00:11:39].

Why Build an Internal AI Team?

The traditional separation between AI model optimization and model serving phases is disappearing in GenAI systems [00:10:29]. In GenAI, everyone acts as an engineer who can optimize overall system performance, creating new challenges for tooling and best practices [00:10:34]. GenAI systems, or agent systems, are considered “compound AI systems” which tackle tasks using multiple interacting components, bridging the gap between AI engineers and product engineers [00:10:49]. A dedicated platform team is critical to bridge the skill gaps between these two groups of engineers [00:12:25].

Ideal Candidate Profile

An ideal candidate for a GenAI team is a strong software engineer who can build infrastructure integrations [00:11:58]. They should possess good developer product management (PM) skills to design interfaces [00:12:07]. Ideally, they would also have an AI and data science background to understand the latest techniques [00:12:14]. These individuals are capable of learning the latest techniques while remaining hands-on [00:12:19].

However, finding such a candidate is rare [00:12:25].

Hiring Principles

When hiring for GenAI teams, several principles can guide the process given the difficulty in finding unicorn candidates [00:12:35]:

  • Prioritize Software Engineering Skills: Strong software engineering skills are generally prioritized over AI expertise [00:12:47].
  • Hire for Potential: Due to the rapid evolution of the field, it’s more effective to hire for potential rather than just experience or degrees, as much experience can quickly become outdated [00:13:03].
  • Emphasize Critical Thinking: A constant in this rapidly changing field is that solutions built today might be outdated in less than six months [00:14:14]. Teams must consistently evaluate new open-source packages, engage with vendors, and proactively deprecate their solutions [00:14:21].

Team Composition

Instead of seeking a single engineer with all desired qualifications, a diversified team approach is recommended [00:13:15]. A diversified team might include:

When these diverse skill sets are brought together on a project, strong engineers tend to pick up new skills through collaboration and grow into more ideal candidates over time [00:13:50].