My name is Aditya Mangalampalli and I am currently a first year at UC Berkeley in Berkeley, CA triple majoring in Computer Science, Applied Mathematics, and Business. I am currently pursuing my interests in Machine Learning and Software Design through unique experiences in both research and corporate settings. Feel free to contact me below for further details!
This project involved me working with Python to create a neural network
that could identify over 104 different diseases and conditions based on retinal
scans. I has used OpenCV, Pandas, Tensorflow, and Keras to train this neural network
and to test it. This project ended up outperforming various implementations from
other research institutions. Source code is public and is available here:
Github
This project took advantage of OpenCV and dlib to create a Python
application that was able to track eye-movements and compare them with different
levels of accuracy when tracking a specific item. It was later the precursor for my
website that was published a few months later. Source code can be found
below:
Github
As the captain of First Tech Challenge team 9614 Hyperion, I spend my
time creating different vision algorithms for different challenges. One year was to
detect blocks, another year was to identify the number of rings in a stack, and this
most recent year was to track a duck. I mainly utilize OpenCV, although sometimes I
also use Keras and other Machine Learning frameworks to create more robust systems.
All my work is public and will be available at the links below:
Skystone 2019-20 Ultimate Goal 2020-21 Freight Frenzy 2021-22
As the Captain of team 9614, I ended up gaining familiarity with
Java-Virtual-Machine (JVM) based languages such as Java, Kotlin, and most recently,
Scala. I work with my team to design various software components that are utilized
throughout the various challenges. I try to incorporate C++ and other embedded
languages to speed up various computationally intensive processes utilizing the Java
Native Interface (JNI). Our code is all public and can be found below:
Github
I conducted research with local professors to collect data from patients
in regards to their eye-movement and any reflective neurological conditions. This
project involved ReactJS, Next.js, Python to train the model and tf.js to keep the
model embeddable and running in a web browser. The website is currently public and
may be somewhat buggy, however the source code is not public due to the fact that it
is currently under development. Check back to see if it is public!
Website Github
I have recently gotten started working on manipulating time-series data
to create algorithms that can predict and forecast events through high-performance
Recurrent Neural Networks. Most of these networks are designed using custom
architectures and have been able to perform at a high level thus far. Currently
source code is unavailable due to the fact that it is currently under development
but should be public within the next few weeks. Keep checking to see if the source
code is available!
Github