Working with AI involves more than just creating code.
Take 25-year-old Pranjali Ajay Parse, an Autodesk data scientist, for example. Her project involves creating an artificial intelligence application that gives workers insights into their work habits, including meeting patterns and work schedules.
Parse added that after working at Autodesk for more than a year and earning her master’s degree in computer science, she has a better understanding of what it’s like to work in an AI role, which is different from what most people may think.
Parse stated that although working in AI may involve technology, the field also places a strong emphasis on ethics. AI work is also primarily interdisciplinary and collaborative. She dispelled some of the fallacies around AI roles in an interview.
It’s not just coding
According to Pranjali, having Python skills won’t get you a job in artificial intelligence.
According to Parse, applicants don’t always require an AI degree to work in the industry. However, she stated that you must be proficient in coding, SQL queries, and case study analysis. She advised candidates to pursue personal projects or boot camps to hone their skills in such areas.
AI is interdisciplinary by nature, according to Parse. It incorporates concepts from a number of fields, such as statistics, computer science, mathematics, and domain-specific expertise.
Parse stated that data science, which necessitates examining and interpreting data sets, makes up roughly 70% of her work. She claimed that software engineering, pipeline construction, data engineering, architectural design, and a lot of arithmetic occupy the remainder of her time.
Parse also mentioned that because technology is always changing, it’s critical to keep up with developments in adjacent fields.
AI roles are often highly collaborative
Though it’s not uncommon for software developers to be loners, working in artificial intelligence doesn’t guarantee solitude.
Parse stated that although certain engineering positions are often autonomous, “AI projects are rarely done solo.” She explained that part of the reason for this is that AI is a new technology that calls for cooperation between numerous teams and stakeholders.
To construct an AI recommendation system project, for instance, Parse stated she must communicate with seven or eight teams.
According to her observations, a data analysis team gathers and prepares the data at the start of the process. Following that, data scientists use modeling and statistical techniques. After that, the machine learning team develops and improves the model. Software engineers create the front end after UX and UI specialists have finished designing the user interface for the model.
The product launch approach was ultimately decided by the marketing team. Parse stated that extensive communication and teamwork are necessary for an end-to-end AI project.
You should be considering ethics
When sensitive data is handled during AI development, privacy teams are frequently deeply ingrained in the process.
According to Parse, there are many privacy protocols. Employees must obtain consent before beginning any task involving a person’s data. In addition, projects need to implement strong production controls, such as pseudonymizing identities and making sure models don’t “unintentionally reproduce biases or create inequitable outcomes.”
This necessitates following legal and regulatory obligations, she stated. It also entails considering the long-term repercussions of undertakings, such as unforeseen consequences and ethical quandaries.
While privacy may appear to be an obvious worry for people working with AI, Parse stated that it is easy to get caught up in how the models perform. Also, because so many teams contribute to the product, it’s easy to focus on your own work rather than the big picture, she said.
According to Parse, it is the responsibility of companies to train their personnel on adequate privacy and ethical norms. Employees should also take a third-person view on their job.
