In between professional commitments and the craziness of life, I love diving deep into personal projects. They’re my playground – spaces where I experiment, innovate, and get to just create something super cool! Below are some of the creations I'm particularly proud of. Let's take a journey through them.
Enter the world of "Pied Poker," a sophisticated poker probability engine I meticulously crafted from
scratch. Driven by the powerful Monte-Carlo simulations, it's not just another poker calculator — it's a
game-changer. The tool delves deep, calculating probabilities that would traditionally be computationally
intensive or downright impossible to derive in real time.
Yet, with optimized hand calculations and nimble simulation performance, "Pied Poker" delivers these statistics at lightning speed, all within seconds. Beyond its capabilities, I took it a step further by launching the engine as a Python package on PyPI. Now, not only can I boast about its prowess, but developers and poker enthusiasts worldwide can seamlessly integrate and benefit from it. With hundreds of downloads each month, this project stands as a testament to the blend of my love for data, coding, and the intricate game of poker.
Check out the Google Colab notebook I created and play around with Pied Poker for yourself!
Meet "Baus Playlist Maker," my brainchild that blends the realms of music and machine learning.
Have you ever heard a song and wished you had an entire playlist emanating the same vibe?
That’s the premise I started with. By engineering a unique machine learning model, this app takes in a
single "seed" song and, like a musical alchemist, curates a playlist echoing its essence.
The challenges? Navigating the curse of dimensionality (with data spanning across 13 dimensions) and optimizing the runtime. But with a blend of innovation and tenacity, I was able to leap over these hurdles, taking the efficiency from Θ(n) to a slick Θ(log(n)). This allows users to create unique and incredible playlists in a matter of a split-second.
The app, now available on the iOS App Store, seamlessly integrates with Spotify API, ensuring users are just a tap away from their next favorite playlist. With a database enriched with over 800,000 songs, each featuring quantified attributes, "Baus Playlist Maker" isn’t just an app; it's a musical journey tailored for every user.
Navigating the bustling world of sneaker culture, I realized there's an undeniable need: distinguishing
the genuine from the counterfeit. The counterfeits had gotten so close to the real-thing, that even
experts were often unable to distinguish real from fake. That's when "Authentic8r" was born. It's not
just an app; it's a fusion of my sneaker passion with cutting-edge tech. Delving deep into the intricate
details of sneaker designs, I built a robust convolutional neural network using Tensorflow to
meticulously analyze and predict the authenticity of a sneaker.
Manually gathering and labeling a whopping 12,000 images was not the most fun I've ever had, but it allowed me to train my model and achieve an impressive 85% validation accuracy. This might sound all tech, but the real magic? Giving users an instant verdict on their sneakers' legitimacy right from their phones. With thousands of verified users and a continuous influx of images aiding model refinement, "Authentic8r" stands as a testament to the potential of machine learning in the everyday world.
Since launching the "Authentic8r" app in 2018, I have attainer over 2,000 users and have collected over 20,000 additional self-labeled images from users, enabling me to improve the accuracy of the model further. "Authentic8r" is available on the iOS App Store.