Suddenly, it seems like everyone wants to be an AI developer, but only a small minority of people have the capabilities needed to actually design, develop, and implement enterprise-grade AI models and systems.
Are you looking for AI staff or consultants? Here are 8 attributes to look for in in-house or external AI developer candidates through email interviews with experts.
1. Get a thorough understanding of the basics of AI
Sunil Kalra, head of data engineering at data analytics company LatentView Analytics, says AI developers should at a minimum understand generative AI and large language models (LLMs). He lists other foundational concepts as UX design principles, security and risk management, LLM lifecycle management, and LLM fine-tuning. Familiarity with vector databases and their governance is also important, but experience with multiple generative AI service offerings will add value. “In addition, business knowledge is essential for effective agile engineering because it allows for the translation of business objectives and domain-specific expertise,” Kalra says. “This combination of technology and business expertise is what makes for successful AI implementations.”
2. Programming ability
Candidate skills should include programming ability in languages such as Python and R, expertise in machine learning algorithms, analytical abilities, domain knowledge, and problem-solving skills, advises Kenny Brown, managing director at business advisory firm Deloitte Consulting. The combination of domain knowledge and a high awareness of unconscious bias is crucial for AI developers, especially in today's context. The old adage “garbage in, garbage out” still applies, albeit in a more subtle form, and guardrails must be implemented to mitigate unintended consequences.
3. Strong system knowledge
Ensuring that AI developers have a strong foundational knowledge of AI systems and the frameworks they need to be built on is key, notes Eric Velte, CTO of ASRC Federal, a consulting firm for civilian, defense, and intelligence agencies. “AI development is different from traditional software development, and it requires specialized skills to get the full benefit,” he explains. Velte warns that current technology makes it too easy to trust models and their results. “This is a danger that needs to be overcome with statistical methods and healthy skepticism.” Hiring developers with system knowledge not only improves efficiency, but also helps secure an organization's data.
4. Approach to data management
Andrew Fedorchek, CTO of Technology, Data and Analytics at Mastercard, said that in addition to technical skills, candidates should have strong data stewardship qualities. Stewardship includes “ensuring security and privacy, and maintaining transparency and control over data use.” It also means “embracing diversity for inclusive and equitable outcomes, maintaining integrity to minimize bias and unexpected consequences, fostering innovation to increase the benefits of data use, and leveraging data for positive social impact.”
5. Strong belief in AI ethics
AI developers need skills that go far beyond technical proficiency, notes Nick Elsberry, leader of software technology consulting at digital transformation specialist Xebia. “In the early stages of widespread adoption of AI, hiring talent that is committed to ethical practices is paramount,” he adds. Developers must also work toward sustainable efficiency to ensure AI systems are developed and deployed responsibly, while protecting against privacy violations, bias, plagiarism, and other common hazards. Elsberry recommends that AI developers also follow corporate ethics and governance practices.
6. Mastering Mathematics and Statistics
Nate Dow, technical director at IT services company BairesDev, says understanding concepts like linear algebra, probability, calculus and statistics is crucial for developing algorithms, analyzing data and building models. Knowledge of machine learning and deep learning is also important, as is familiarity with various machine learning algorithms.
7. Strong data management skills
Ever since the “big data” movement began several years ago, many IT leaders have recognized that much of their work is about collecting and organizing data. Mike Loukides, vice president of emerging technology content at educational publishing and services company O'Reilly Media, says AI won't change that perspective. It's not a matter of taking a foundational model that incorporates all the knowledge in the world and applying it to any problem at will, he says. “You need to collect data to fine-tune,” he says. AI professionals need to understand what that data means and what biases are built into it.
8. Excellent communication skills
Beyond technical skills, candidates must be able to clearly and concisely explain concepts and results to business stakeholders, advises Jayaprakash Nair, head of analytics at data and digital engineering services company Altimetrik. “AI explainability is a big challenge for the industry,” he notes. Business leaders must not be confused or misled by arcane terminology and jargon.
