Must-Have Skills for AI Success: Top AI Advocate Breaks it Down

I talked to Apoorva, a Senior AI/ML Developer Advocate at MongoDB, about her journey into the field of Artificial Intelligence and Machine Learning (AI/ML). Apoorva shared her experiences working as a Data Scientist and now an AI Developer Advocate.

‍

Q: How did you get into AI/ML? Tell us about your journey.

A: It all started way back in high school! My first encounter with coding was through a Computer Applications class. Let's just say it involved a lot of LeetCode-style problems, which were quite intimidating at the time. This initial experience actually scared me away from computer science as a major. I ended up pursuing electrical engineering and even started a Master’s degree in it, but I finally realized that it wasn't working for me.

That's when the buzz around machine learning was starting to gain momentum. Curiosity took over, and I decided to take a machine learning course. It was a turning point. I was coding in Python, a much friendlier language, and I was hooked. I spent the remainder of my master’s program doing a thesis in natural language processing. This hands-on experience solidified my interest in pursuing a career in AI/ML.

However, landing that first job wasn't easy. Despite having a Master's degree, I sent hundreds of applications and even resorted to cold emailing directors. Persistence paid off! After nearly 400 applications, a security company took a chance on me, and that's how I embarked on my journey as an AI/ML engineer specializing in cybersecurity.

‍

Q: You have a Master's in Computer Engineering. Can you elaborate on the differences between MS and PhD in the context of pursuing a career in AI/ML? Is a Master's degree sufficient for most roles?

A: In my experience, a Master's degree is sufficient for many AI/ML roles, especially those focused on applying machine learning to solve specific problems. Having a solid foundation in core machine learning and statistical concepts, good engineering skills, and a general inclination for research can take you far.

However, if your goal is to delve into cutting-edge research and develop entirely new algorithms or foundational models, then a PhD might be useful. The deeper mathematical background that comes with a PhD can be advantageous in such research-oriented roles.

Ultimately, the choice depends on your career aspirations. Do you want to be on the application side, solving problems with existing techniques, or do you see yourself pushing the boundaries through research?

‍

Q: You were a Research Scientist building machine-learning models at FireEye. In your experience, what are the key differences between the roles of a Research Scientist and an ML Engineer?

A: Here's my take: Data Scientists and Research Scientists typically focus on training models, prototyping, or A/B testing and experimentation if you are on the Product Data Science side. Machine Learning Engineers, on the other hand, lean towards building and maintaining the infrastructure that allows data scientists to efficiently train their models.

The specific responsibilities can also depend on the size of your team as well. In smaller teams, you might find yourself doing a bit of everything, encompassing the full data science and machine learning engineering stack.

‍

Q: What are the typical tasks and responsibilities of a Data Scientist? Can you share an example of a project you worked on as a Data Scientist?

A: Data Scientists wear many hats! They're involved in the entire data science lifecycle, from data collection and cleaning to model training, evaluation, and deployment.

One of the most interesting projects I worked on as a Data Scientist at Elasticsearch involved entity scoring. In cybersecurity, entities can be anything from users and laptops to servers and databases. The goal was to develop a system that assigned risk scores to these entities based on their behavioral patterns to help security analysts easily find malicious entities within an organization for investigation.

‍

Q: This project sounds like it required specific cybersecurity knowledge. How did you transition into cybersecurity without a background in the field?

A: My foray into cybersecurity was somewhat accidental. It stemmed from applying to a wide range of positions, and one of the references I had just happened to be at a cybersecurity company. The beauty of cybersecurity is that it's accessible to people from diverse backgrounds. There are always subject matter experts who can provide the necessary domain knowledge to be successful.

‍

Q: What do you do as a Senior AI/ML Developer Advocate? Can you walk us through a typical day in your current role? 

A: My current role as a Senior AI/ML Developer Advocate allows me to combine my experience as a Data Scientist with my passion for education. I essentially act as a bridge between AI/ML developers and the technology my company offers, MongoDB. This involves creating educational content, such as written guides or hands-on workshops, to help developers leverage MongoDB effectively in their AI/ML applications.

A typical day can involve a variety of tasks, from researching and planning content to developing workshops and delivering presentations. I also get to interact with the developer community through various channels, answering questions and providing guidance.

‍

Q: What skills and qualities are essential for success in AI/ML roles?

A: Adaptability is key in today's rapidly evolving AI/ML landscape. The tools and techniques we use today might be obsolete in a few months. Embracing change and being a lifelong learner are key qualities for success.

Another interesting trend is the merging of roles. Traditionally, we had data scientists, machine learning engineers, and software engineers. Now, we're seeing the emergence of the "AI engineer" who can leverage AI as a versatile tool across their daily tasks. This ability to integrate AI into various aspects of software development will be increasingly valuable.

‍

Q: What advice would you give to people interested in AI/ML but anxious about the constant learning curve?

A: A small but crucial step is to start small. Feeling overwhelmed by the vastness of AI/ML is natural. Instead, try incorporating these tools into manageable projects. For instance, if you're typically used to making  Google searches, experiment with AI-powered search engines like Perplexity or ChatGPT. As you gain confidence interacting with these models, you can gradually progress towards more complex tasks that involve coding or automation tools like Copilot.

Remember, the key is to break down the learning process into manageable chunks. Focus on taking the next step, not on mastering everything at once. Every little bit of progress adds up!

‍

Summary

Apoorva’s advice for aspiring AI/ML professionals:

• Start small: Don't be intimidated by the vastness of AI/ML. Begin with user-friendly tools and progress gradually.
• Consider your career goals: Do you see yourself developing new algorithms (research) or solving problems with existing techniques (application)?
• Be adaptable: The field is constantly evolving, so embrace lifelong learning.
• Learn from others: Domain knowledge can be acquired on the job.
• Break down learning into chunks: Don't try to master everything at once. Focus on achievable steps for steady progress.
• Experiment with user-friendly AI tools: Utilize existing AI-powered search engines or automation tools like Copilot to gain confidence.

‍