Discover more from DSBoost
1 model - $2M+ revenue. We interviewed the man behind. - DSBoost #24
💬 Interview of the week
This week we interviewed Harpreet, who is a DevRel Manager. Enjoy:
What did you study/are you studying (if your background is different from DS, how did you end up in the field)?
During my undergraduate years, I focused on Econometrics and graduated in 2007. Unfortunately, that coincided with a recession, leading to a scarcity of job opportunities.
Honestly, I could have been a better student during my undergrad days. I barely made it through, had no internships, and lacked significant skills.
As fate would have it, I taught high school mathematics for two years.
Surprisingly, during this experience, I discovered my deep passion for math. I had always been good at it as a student, but teaching and communicating it to others made me realize how much I loved it.
This realization sparked my interest in math-related careers, leading me to stumble upon Actuarial sciences. However, becoming an Actuary meant passing numerous exams that required advanced math skills.
Naturally, I believed that pursuing graduate school was the path forward. Unfortunately, my less-than-stellar grades during my undergraduate years made it challenging to gain admission to any graduate programs.
But hey, I wasn't about to give up.
I started from scratch and took mathematics courses at a local community college in the summer of 2009. I began with pre-calculus and gradually worked through fundamental calculus and linear algebra over three semesters. I then continued my studies at UC Davis, where I delved into more advanced statistics courses.
Finally, in spring 2011, I completed all the prerequisites for a master's degree in mathematics, achieving a 3.80 GPA and scoring in the 97th percentile for the math portion of the GRE.
That marked a turning point for me.
I began my career as an Actuary in 2013, but life had more surprises. I eventually shifted gears and entered the field of biostatistics, exploring new horizons. Then, in 2019, I landed my first official data science job as a senior data scientist. It was quite a milestone, considering my journey up until then.
But wait, there's more!
Later that same year, another company recognized my skills and offered me the lead data scientist role. It was an exciting leap forward in my career.
In 2020, I ventured into podcasting and launched "The Artists of Data Science" podcast. To my amazement, it caught the attention of Comet, who sponsored my show for an entire year. In August 2021, Comet took it a step further and offered me a position as a Developer Advocate.
And now, here I am two years later, thriving in Developer Relations.
There were moments of stress, doubts, and uncertainty along the way. That pivot into data science was one of the most challenging times in my life. I felt pathetic; I was at 35 years old, still trying to find my way and chart my course. I felt like I had all the pieces to the puzzle, but I couldn't connect them. Every job I applied to, I was steadily just getting rejected. But I never once let the setbacks I faced make me question if I was out of my mind or wasn't cut out for it.
Despite all that, you can chart a remarkable path with perseverance and a genuine passion for what you want. Life truly is an unpredictable adventure!
What are your favorite resource sites and books (ML/AI)?
Deep Learning Illustrated: A Visual, Interactive Guide to Artificial Intelligence Book by Jon Krohn
Deep Learning: A Visual Approach by Andrew Glassner.
Why these two books?
While I'm not an engineer or researcher professionally, I like to understand things better by taking a step back and looking at the bigger picture first.
I'm not afraid to get into the nitty-gritty details, especially since I have a strong math background.
But it's essential to focus on intuitive understanding before diving into the math.
In this field, it's important to keep up with the latest developments and only go in-depth when necessary. If you've got a good grasp of math, up to Linear Algebra and Differential Equations at the undergraduate level, you should be able to understand research papers easily.
What do you enjoy the most in your work?
I love being in DevRel, especially in the ML/AI space.
It's like I'm in this amazing tech playground where I play and learn every day. It's so cool to get hands-on with the latest libraries and tools, then break them down into digestible bits for others through my tutorials and guides.
But you know what's even better?
The community vibe. Building a space where people can learn, share, and geek over the same things I do? It doesn't get better than that.
For me, it's all about the tech, the learning, and the people.
That's why I love my job.
Do you use ChatGPT or other Al tools during your work? If so, how do they help you? Do they change your approach to problems?
I use ChatGPT daily for various purposes, such as generating boilerplate code, assisting me in writing and research, providing therapy sessions, career coaching, and so much more.
What is your favorite topic within the field?
Deep Learning, specifically generative models and foundation models
Which one of the recent AI/ML models will have the most significant impact on the industry in your opinion?
What are you currently learning or improving (topics you are interested in nowadays)?
I've always been drawn to deep learning for it's ability to augment human creativity. So I've been especially excited with the advancements we've seen in Generative AI, Language Models, Foundational Models, and all the rest.
I'm most interested in how to use these new models, interact with them, chain them together, and build agents to help make my life easier.
What is your most significant achievement?
My most significant achievement in data science was deploying my first model to production, which brought in $2 million in revenue during the first quarter of its launch.
I took charge of the entire machine learning project, working closely with the end users to understand how they made decisions. By diving into the raw data and their interpretations, I crafted a robust set of features through smart feature engineering.
I experimented with different models and features, eventually finding the one that accurately represented the data. Along the way, I learned the value of managing experiments effectively.
To ensure performance, I set up a model monitoring system that tracked deviations between predictions and user actions, introducing me to MLOps.
Lastly, I collaborated with an engineer to deploy the model successfully in production.
I got to flex my technical skills while learning the importance of collaboration, feature engineering, experiment management, and MLOps in delivering impactful data science solutions.
Can you share a fun fact about yourself?
tl;dr: I'm a hustler baby, I just want you to know.
I turned the big 4-0 this year and parenthood has become a central part of my world, with a lively three-year-old and an adorable 7-month-old keeping me on my toes.
As I navigate this phase, I'm still trying to find my footing in my career.
It's a journey of self-discovery; truth be told, I often question my abilities and battle imposter syndrome daily.
But I've always been a hustler. It's in my DNA.
I do what I got to do to get to where I need to be. I've been putting in the work my entire life, doing whatever it takes to inch closer to my goals.
I'm still piecing everything together, but one thing's for sure: I've developed an unyielding resolve.
Nothing slows me down anymore.
I keep pushing forward, undeterred, because I know I have what it takes to handle whatever life throws my way.
So, let's keep it real.
I'm not claiming to have it all figured out or be some untouchable superhero.
I'm just a regular person trying to make things work.
But what sets me apart is my unwavering determination and resilience.
Despite the setbacks and doubts, I keep pushing forward, knowing that with each step, I'm getting closer to where I want to be.
And that's what makes this journey worthwhile.
Thanks for reading DSBoost! Subscribe for free to receive new posts and support my work.
🧵 Featured threads
🤖 What happened this week?
A long interview with DeepMind CEO Demis Hassabis about AI-related topics, including organizational decisions, competition, AI risk, and the effects of AI on the labor market. And most importantly: How Google plans to stay competitive.
OpenAI is introducing a new team that will be jointly led by Ilya Sutskever and Jan Leike. The team's primary objective is to pursue scientific and technological innovations that enable the effective management and regulation of artificial intelligence systems that surpass human intelligence.