4 Years Later: Was Flatiron School Worth it for Sr. AI Engineer Varun?

Four years ago, Varun Kumar was set to become a psychiatrist when he made a career pivot into software engineering. Relying on his med school grit and determination, Varun graduated from Flatiron School and then became a Software Engineer at IBM. Now Varun is a Senior AI Engineer and reflects on the important lessons he’s learned from his career change and what it takes to break into artificial intelligence. Four years later, does Varun think that Flatiron School was worth it?

What inspired you to pivot from working in medicine to working in tech?

I wanted to be a psychiatrist, so I was focused on psychology, biology, chemistry, and physics. I really enjoyed the classes, but I didn't want to work in a hospital — I wanted to have the ability to build solutions for people on a 1:1 scale. As I was working towards my degree, I realized there was a set capacity for the amount of people I could actually affect, and I realized that statistically there was a 1-2% chance of becoming a psychiatrist after medical school. I took a step back and considered what other options were available to me. 

A great friend of mine was a software engineer and I explained to him my predicament and he suggested applying to business roles. Rather than thinking about what I can do with my psych/neuro background, I started asking myself what I want to do and worked backwards. My friend pulled up a coding problem and told me to tell him how I was going to solve it. I didn’t know how to code at all, but he told me to share my ideas and he’d code it — this experience totally opened the doors for me! I solved my first coding problem and it wasn't because I knew how to code — it was because I had someone who showed me that it was possible for someone without a computer science background to code. My friend was adamant: Don't worry about what you're doing, just code every single day. So, I coded every day until I got into Flatiron School, like it was a job. I'm naturally introverted and I was so drained by every single day I talked to patients, but I’d go home tired and still have the energy to code. 

There are so many coding bootcamps now — What stood out about Flatiron School?

I considered several popular coding bootcamps, curious to see if graduates enjoyed their time there and got what they wanted out of the program. I reached out to bootcamp alums about their experiences. I wanted their earnest opinions of the program, more than rates and stats. Flatiron School alumni had great things to say about the bootcamp!

What did you learn in the Software Engineering bootcamp?

We started with simple terminal programs, learning Ruby, JavaScript, React, CSS and Git. I learned how to think about a project, how to change careers, network, advocate for myself to stand out in the hiring pool, as well as how to interview and present demos to people who don’t understand code. I also learned how to collaborate with others and how to debug a problem. When I left Flatiron School, I was confident that I could solve any problem with enough time and effort. 

How did Flatiron School prepare you for the tech job hunt? 

Flatiron holds Career Week events, where I had the opportunity to get paired directly with an employer. This was important because I got my first offer through that experience! It boosted my confidence that I can land a software engineering position and that I don’t need to keep going to school to show people I was capable. I got more out of networking than personal branding. I put myself out there and was physically trying to be in front of as many people in the engineering world as possible. 

What was the first tech role you landed after Flatiron School?

My first tech role after graduating from Flatiron School was at a real estate company that wanted to build a platform where they could include their listings so that agents could interact with them. I built the platform in React and was able to get a taste of the startup world. I was the only person on the tech team — just me, the CEO, and the realtors. It was too much for a junior, so I left and got a job at IBM for three years where I worked on front end, then moved to full stack, then back end AI work. 

Was IBM interested in your Flatiron School bootcamp experience?

Absolutely. I'm really good at demoing so I started by sharing a project with them that I made around bootcamp analytics. It was a combination of luck and skill, but IBM was definitely interested in my bootcamp experience.

What kinds of projects did you work on at IBM? 

I started by working on the front end of the application in security. We visualized data to make sure people who are working at the company know when a user is doing something nefarious, even unintentionally. 

Eventually, I started getting opportunities to lead my own team. I was the lead developer and scrum master on a project called the Cyber Range Portal, which was a virtual hacking space where you could log in, get hacked, and then respond to it in real time. If you did it wrong, you’d get feedback and lessons to help you improve. The project was aimed at non-technical people, but it gave me a taste for what it felt like to have my own project and work with a smaller team. 

When you're working for an enterprise company, you can be working with 50 people — the project becomes more important than any individual's work no matter how hard they try. 

Have you had to get any cybersecurity certifications to work on these IBM projects?

I did some cyber training, but I didn't have to get security clearance. I worked a lot on building the actual applications that security professionals use, rather than becoming a security professional.

Now you’re a Senior Software Engineer at a startup! What AI tools are you using on the job?

My full job title is Senior Software Engineer - Python, AI, GIS Mapping. I use Pinecone.io, OpenAI, Langchain, a lot of Python, and Postgres. For the front end, we’re still using React. 

We’re all seeing more job postings with “AI” in the title – what does your job actually look like in AI? What kinds of projects are you working on?

The main innovation in AI that has occurred is the invention of an architecture known as the Transformer model. It is my job to use different models to eventually generate something useful, from a natural language prompt.

If I break it down further it looks like this:

• Research: Stay updated with the latest advancements in the field. This could involve reading research papers, attending seminars or webinars, and participating in online forums and communities. This is crucial as the field of AI and machine learning is evolving rapidly.
• Data Preparation: Work on preparing and pre-processing the data for training language models. This involves collecting data, cleaning it, and converting it into a format that can be used for machine learning.
• Model Development and Training: Design and implement machine learning models. This includes choosing the right algorithms, tuning parameters, and training the model on the prepared data. This process often requires running experiments and making iterative improvements based on the results. Many times, I am building on pre-trained models with either fine tuning, or instruction via prompts.
• Model Evaluation: Evaluate the performance of the models using appropriate metrics. This often involves testing the model on a held-out validation set and analyzing the results.
• Collaboration: Work closely with other teams, such as product development, to integrate the AI models into products or services. This could involve optimizing the model for deployment, working on the user interface, or addressing user feedback.
• Documentation and Presentation: Document the work for future reference and present findings to stakeholders or to the technical team. This might involve writing technical reports, creating presentations, or showing working code.

Are there any specific prerequisites that someone needs to get a job in AI?  

I don't think so! This field is moving fast and if you are going to be technical you basically need to show that you know enough to be dangerous in whatever task the company is trying to build a solution for. 

I think the prerequisites really depend on the company mission, their size, and their budget. If you are looking to work for a research heavy organization like OpenAI, Google Brain, or meta AI, publications and academic NLP knowledge may be required. But many other businesses are working on integrating the recent AI advancements internally, where you might be able to showcase what you have independently built as evidence that you are hireable.

So far, what are some lessons you’ve learned in your tech career journey? 

1. I think it's important to switch jobs every so often if you can, especially in tech since it’s always changing. 
2. You can't have the mindset that after a bootcamp you’re finished learning. No matter who you are or how senior you get, you will have to learn again. As long as you keep doing that, you'll be a great engineer. 
3. You have to look out for yourself as a technologist. Often, people will tell you things look good when you know they’re not. Stay vigilant with your skills. 
4. If you’re interested in a tech career, you can do it. You don't have to be a genius, you just have to want to! 
5. You have to like being in front of a computer. 

At this point in your tech career, was Flatiron School worth it for you? 

For me, this career change was absolutely worth it — with a huge caveat because I definitely got lucky. You get what you put into it. I definitely put in a lot of effort while I was doing it, and I didn't just rely on the Flatiron School materials to guide me. I did a lot of outside research and talked to engineers as soon as possible. 

If you think you want to be a developer, do an immersive program! I don't think you should do it part-time or just take classes from Udemy. Most of the time what ends up working is an immersive program. If you're super motivated and self-directed, you can do an immersive program on your own, but the majority of people (me included!) need to be around other people who are all coding.

Find out more and read Flatiron School reviews on Course Report. This article was produced by the Course Report team in partnership with Flatiron School.