AI Resilience: Upskilling in an AI Dominant Environment AI Resilience: Upskilling in an AI Dominant Environment
At this year’s ODSC East, Leondra Gonzalez gave a compelling talk on navigating a career in data science given recent technological... AI Resilience: Upskilling in an AI Dominant Environment

At this year’s ODSC East, Leondra Gonzalez gave a compelling talk on navigating a career in data science given recent technological advances in AI. Since ChatGPT became widely available to the public in 2022, AI technology has become even more buzzworthy. Advances in large language models and other techniques’ ability to process huge amounts of unstructured data have changed the game in a variety of domains; data science is no different.

Data science is a diverse field, encompassing disciplines of statistics, programming, mathematics, business intelligence, and computer science, among others. No one can know everything, and each role requires slightly different skills, so data scientist positions tend to require more expertise in some disciplines than others. Leondra mentioned that she has noticed a shift in the expectations for data scientist roles. Roles that previously didn’t require much natural language processing (NLP) or multimodal modeling, now may incorporate those techniques as key to some aspect of the position. Even positions that don’t focus on AI technology explicitly are experiencing change as some tasks are expedited or automated. For example, writing basic SQL queries is a task that LLMs can easily tackle with enough knowledge of the database’s schema. As the data science field evolves over time, staying current with the latest technology is a necessity for individual contributors in data science.

For data scientists who want to ensure their skill set remains relevant over time, Leondra recommended having a pointed, short-term goal for one’s career and monitoring how the technology has evolved for that next step. This involves researching the expectations of positions outside one’s own current position or discipline to understand how the market is changing more broadly. Job descriptions can indicate how data scientist positions are changing, however, Leondra pointed out that they typically contain frameworks or buzzwords rather than the business problems they expect the role to solve. She mentioned that the best candidates are able to read between the lines in job descriptions and anticipate the business objectives among the noise of frameworks and packages. From the hiring manager’s perspective, Leondra recommended looking for candidates with strong work ethics and foundational skills, and to support an upskilling culture.

Acquiring new skills that fit the changing expectations can be a time-consuming, but rewarding experience. The amount of work involved with taking courses and reading papers on top of existing responsibilities can be a bit daunting. Being efficient in identifying top priorities to learn and the most easily consumable source of information can lighten the workload. Leondra emphasized that data scientists should focus on learning new fundamentals rather than frameworks. For example, it would be better to understand the core techniques behind LLMs rather than being very fluent in LangChain. Mastering the foundations enables one to be more flexible and deliver value more easily to a new role. Key concepts that were noted as increasingly relevant include LLMs, embeddings, variational autoencoders, transfer learning, and transformers. Individuals may consider starting to learn new concepts that are adjacent to one’s existing expertise rather than jumping into completely new disciplines right away.

After acquiring foundational knowledge of relevant concepts, it is most impactful to solidify one’s experience by applying them in real-world scenarios. Because the skills one learns are likely to be new to the team or company, there is often low-hanging fruit in the form of previously unsolvable business problems. Leondra recommended looking for opportunities to solve such business problems in one’s current role to complete a round of the upskilling cycle.

One of the best ways to keep a pulse on the changing job landscape is to take advantage of networking events. This year’s ODSC East in Boston hosted a number of well-attended networking opportunities, in addition to having direct access to speakers after their sessions. If you missed this year’s eastern event, check out ODSC Europe September 5th-6th or ODSC West October 29th-31st!

More Links:

Our recent podcast interview with Leondra

Watch this session for free on Ai+ Training

See the full slide deck here

Nathaniel Jermain

Nathaniel is a senior data scientist in the marketing industry, located in Saint Petersburg, FL. The focus of his work includes machine learning, statistical analysis, and a particular interest in causal inference. Feel free to connect with Nathaniel on LinkedIn: https://www.linkedin.com/in/njermain/