hey there 👋
hey, I’m Dev, and if you’re new to my bi-monthly newsletter, welcome! My bi-monthly newsletter is where I recap what’s been going on in my life and share some thoughts and reflections from the last couple months. Allow me to introduce myself, I’m currently working in Medical Imaging x AI under Dr. Tyrrell; looking to integrate Artificial Intelligence into a clinical setting to improve the diagnosis process. These last couple month have helped me go from ground 0 to 1 and then from 1 → 100. As I’ve made these strides, I’ve understood the importance of discipline and iterative progress. Let’s get into the thick of what I’ve been up to this month, hope you enjoy the newsletter!
iterative progress.
As I’m writing this newsletter on a snowy February day, I’m reflecting on the contrast between me now and myself a year ago. As I reflect on my personal growth journey, I realize the importance of acknowledging and analyzing the impact of 'quick' wins. Looking back on my experience from a year ago, I remember feeling accomplished and motivated by the quick wins I was able to achieve, but as time passed, I realized that these wins were not necessarily in alignment with my long-term goals.
I was reading Atomic Habits by James Clear earlier this month and one of his quotes really resonated with me. He said
"You do not rise to the level of your goals. You fall to the level of your systems."
I now understand that true progress is not about achieving quick wins or lofty goals, but about developing systems and habits that will lead me towards sustainable growth and success. It's about making meaningful and consistent changes that align with my values and long-term objectives.
While quick wins can provide a sense of momentum and satisfaction, it's essential to consider whether they are leading us towards our ultimate objectives. In my case, I wasn't actively seeking out quick wins, but they still found their way into my path. And while they gave me a boost of energy, they didn't necessarily contribute to my long-term growth. In hindsight, I now understand that true progress is not about achieving quick wins. It's about making meaningful and sustainable changes that align with our values and long-term goals. Rather than focusing solely on short-term gains, I need to develop habits and behaviors that will lead to sustainable growth and success.
Fast forward to December 2022; staring at my laptop screen, unsure what to do next. Despite changing and growing so much between those months, in some ways, I felt like I'm back where I started - unsure of what to do next. But this time around, I'm armed with something I didn't have before - the knowledge that quick wins and immediate gratification are not the keys to long-term success. Instead, I know that true progress comes from building sustainable habits and systems that support my goals.
So, instead of succumbing to the paralysis of indecision, I remind myself of the habits and systems I've worked so hard to cultivate over the past year. I focus on taking small, intentional steps towards my goals, knowing that each action I take is a step towards the person I want to become.
Around this time, it had been the first time in a while I was building in the AI space - I was definitely rusty. It took some time to get back into the flow of things and start building effectively. I knew that the key to success was to keep taking action and making incremental improvement or iterative progress. So I decided to set a goal for myself - to build a new project every week over the course of the next two months.
Looking back now, 8 projects later, in those moments, I didn’t feel that instant gratification of the ‘quick’ win that I had experienced before. Instead, it felt like I was climbing up one stair on the staircase; each project was a step forward, each new habit I built was a step up the staircase towards my goals. There were days when I felt stuck, when I wasn't sure if I was making progress or if my efforts were in vain. But I continually reminded myself that the process and the lessons learned along the way are just as valuable as the ultimate outcome.
As I kept building, taking small steps every day towards my goals, I realized progress is not always linear - there will be setbacks and obstacles along the way, but it's important to keep moving forward and stay focused on the bigger picture. Now, looking back at those 8 projects, I am proud of the progress I have made. I know that there is still a long way to go, but I am confident that the habits and systems I have built will continue to support me on my journey.
discipline > motivation.
Over the course of the last 2 months, there have been days where getting up hasn’t been as easy. While it's tempting to attribute it to the dreary winter season, I know deep down that it has more to do with my own discipline and motivation levels. I had this talk with one of my past mentors, where it's easy to fall into the trap of relying solely on motivation to get things done, but motivation is fleeting and unreliable. Instead, it's the daily practice of showing up and doing the work that leads to sustainable progress. At that time, the concept conceptually and theoretically made sense in my head, but it was hard to internalize.
As I continued to delve deeper into building projects within the AI space, I realized that relying on motivation alone wasn't sustainable. I needed something more to keep me going even when the initial excitement wore off. That's where discipline came in. I started implementing small habits into my daily routine that would contribute to my larger goal of becoming well-versed in AI. I would set aside specific blocks of time each day for learning and practicing, and I would hold myself accountable to sticking to that schedule. Of course, there were still days where getting up and doing the work was a struggle. But with each small win and each successful habit, I found myself building a foundation of discipline that helped me push through those tough moments.
Building discipline was a rather difficult task, and it took time and effort to cultivate. One of the biggest learnings that I walked away with is that asking the why behind what I’m doing. To find that reason, I had to dig deep into my personal experiences. I reflected on the times when I had succeeded and failed, and what had led to those outcomes. I realized that discipline was often the deciding factor between success and failure. When I was disciplined, I could stay focused and work consistently towards my goals, even when the going got tough. But when I lacked discipline, I would often give up or get distracted, leading to subpar results. While I was actively building projects, I wasn’t driven by the gratification of completing a project and putting it out. Asking questions like “what’s the overarching goal in doing this?” or “why am I doing this?” helped me focus on the bigger picture and the purpose behind my actions. It allowed me to shift my mindset from seeking short-term satisfaction to working towards long-term goals. With each passing day, I realized that building discipline wasn’t just about forcing myself to do things I didn’t want to do. It was about finding meaning and purpose in what I was doing and using that to build a habitual mindset for myself.
Discipline isn’t built overnight, it requires continuous practice - it compounds over time and I noticed as I was building a sense of accomplishment and momentum with each passing day. It was almost like a snowball effect - the more I did, the easier it became to continue doing it. Looking back - building discipline is not a one-time event, it's a lifelong process that requires consistent effort and commitment to the daily actions that align with my goals and values. It's like the art of bonsai, where it's not about creating a beautiful tree, but rather the art of nurturing and pruning to create a living masterpiece - discipline isn't just a choice, but a way of life.
going from 0 → 100 in the AI space.
Over the course of the last 2 months, I’ve spent a lot of time building and iterating on machine learning models. Prior to the beginning of these 2 months, I had spent a ton of time researching and reading papers, but not as much time implementing these papers - I was definitely rusty. I spent some time brainstorming which sub-areas of AI that I wanted to become more knowledgable in. I narrowed this wide list down to Computer Vision and Natural Language Processing.
Initially, it felt like I was hitting a brick wall. I started off with implementing the infamous UNET framework for a segmentation project; I spent countless hours coding and debugging, but at times, it felt like I wasn’t making any progress. One of the key things that I learned throughout this journey is the importance of patience and persistence. Progress was not always linear, and there were moments where I felt like I was stagnating. However, I reminded myself that growth and improvement take time, and that small, consistent steps forward can lead to significant results in the long run. Although I built 8 projects over the course of the last couple months, I’m going to highlight a couple that I enjoyed the most and I’ll hyperlink the GitHub’s for the rest.
diagnosing COVID-19 with CT-scans
One of the first projects that I built was a computer vision project that involved diagnosing COVID-19 based on a CT scan of a patient’s lungs. I personally had a ton of fun building this project as the images were relatively raw and required quite a bit of data pre-processing; using techniques such as resizing, grey-scale, normalization, batch normalization, etc.
This was one of my favourite projects to work on because it was the least straight forward - I ran into problems along the way.
This project forced me to think outside the box and come up with creative solutions to overcome these challenges. It was also the project that pushed me to delve deeper into the theory and fundamentals of AI, which in turn helped me to better understand the practical applications. In more technical terms, I had to re-train my model about 6-7 times because sometimes it would overfit, on other occasions the accuracy wasn’t high enough, etc. I built this model using Tensorflow and trained a Convolutional Neural Network to be able to diagnose Covid-19. My model trained with over 99% accuracy and was able to predict the labels correctly as well. If you want to check it out, here’s the GitHub code:
segmenting nuclei with UNET framework
Another one of the projects that I built out was another computer vision project, but instead of diagnosing/detecting, it was around segmentation, specifically semantic segmentation. I built this project by directly implementing the infamous Convolutional Networks for Biomedical Imaging research paper. Although I enjoyed the previous project quite a bit, this was my first attempt at implementing a paper from scratch and I would spend hours on end just staring at my screen, unsure what to do next. Implementing the segmentation paper taught me something beyond persistence and patience. It taught me the value of breaking down problems into smaller, more manageable pieces. I found that when I tackled the project as a whole, it felt overwhelming and unapproachable. However, when I started breaking the project down into smaller, more digestible chunks, it suddenly felt more feasible. By breaking it down into smaller parts, I was able to make incremental progress, which built my confidence and made it easier to tackle the more challenging parts of the project. Getting into the more technical side of things - I used Tensorflow to implement this paper, but I used other libraries such as SkLearn, split-folders, NumPy, Matplotlib, etc. to pre-process data and perform some data analysis. The model trained at over 96% accuracy and if you want to check it out, here’s the GitHub code:
other projects.
finding my voice in writing.
Over the span of the last year, I’ve done a lot of writing, from papers to essays to newsletters to articles, but I haven’t exactly been proud of the work that I’ve put out. Despite receiving positive feedback and encouragement from others, I struggled to find my voice and felt like I was constantly imitating others instead of developing my own style. It felt like something was missing in my writing, and I wasn't sure what it was. I realized that I had been so focused on crafting a certain image that I had lost sight of the true purpose of writing - to communicate my own unique thoughts and perspectives.
As I reflected on my writing, I realized that finding my voice was akin to finding a unique instrument to play in a band. It's not enough to know how to play an instrument, but to truly make music, you have to find the one that resonates with you. Similarly, with writing, it's not enough to know the mechanics and technicalities; it's about finding a style and voice that is uniquely yours.
With this realization, I started to experiment with different styles, tones, and approaches to see what felt most authentic and natural. I found myself exploring new topics and angles, using more vivid language, and sharing more personal stories. Funnily enough, this newsletter was my first attempt at writing a more reflective piece. Rather than simply summarizing my recent projects, I wanted to delve deeper into what I've learned and share those insights with others. It's a shift in mindset from showcasing accomplishments to offering value through personal experiences and lessons. It's a reminder that writing isn't just about sharing information, but also about connecting with others and sharing a piece of oneself.
I'm still a work in progress, and there's always room for improvement, but I'm excited to continue exploring and honing my writing voice. Just like a musician perfecting their craft, writing is a never-ending journey of growth and evolution, and I'm eager to see where it takes me.
Before ending off this newsletter, I want to share 2 articles that I wrote in the past 2 months → practicing a more formal style of writing.
writing articles.
The articles that I wrote was an attempt at practicing writing in a formal tone. I based my articles on recent papers that have come out in the AI space. It allowed me to take complex concepts and ideas and break them down into more digestible pieces for a broader audience. Through this process, I discovered the importance of writing with clarity and simplicity, without sacrificing the integrity of the content. Writing in this style was like solving a puzzle - trying to find the perfect combination of words to convey my message effectively. It was challenging at times, but also incredibly rewarding when I finally found the right words. I wrote 2 articles, the first one was about leveraging zero-shot learning to diagnose pathologies. The second article I wrote was about detecting COVID-19 with a Triplet Siamese Network. If you want to check them out, I’ve linked them below.
looking ahead.
If you’ve made it this far, I would like to thank you for taking time to read my newsletter. I hope that my insights and experiences have been valuable to you, and I look forward to sharing more of what I’m up to in the future. With that being said, here’s what I’m going to be working on in the next few months:
Building out Biomedical Imaging x AI project - I’ll be working with PhD and Grad students at the lab that I’m researching in to build out a bigger project. More details to come.
Exploring a new subset of AI - I’ve spent the last couple months in NLP and CV, but I think it’s time to shift my focus and building.
Keeping up with writing - I’m going to continue to consistently put out articles and pieces of writing. Ensuring that I don’t lose my voice in writing.
That’s all from me; if you enjoyed reading this newsletter, please consider subscribing and I’ll see you in the next one 😅