Carbon0 is a web app that helps people build carbon negative lifestyles for themselves and their communities.
Technologies: Django, Bootstrap 4, PostgreSQL, Heroku, AWS S3, Chart.js, Django REST Framework, Travis CI
I started headsetsGoodbye to research empathetic technology ideas using Tensorflow.js. We're using deep computer vision to enable intuitive human-machine interaction (GIF source: GIPHY).
Technologies: HTML, CSS, Tensorflow.js, face-api.js, Handsfree.js, and Three.js
I built this while TAing the DS/ML courses at the Dominican University of California. Its purpose was to show our students how deep learning has the power to potentially save millions of dollars (and more importantly, lives) in places like California that experience annual "fire seasons."
Technologies: Python, TensorFlow/Keras, FastAPI, Docker, Heroku
While at Onshape, I built a semi-supervised ML tool to report leading indicators (out of 30+ features) that were correlated to slow performance on our server-side code (for confidentiality reasons, some of the machine learning functionality has been left out of this public repo).
Technologies: Python, Scikit-Learn, Flask, Panda-Profiling
Samaj (or "understanding", as it means in Urdu) is a mini-machine learning framework I started in grad school as a fun side project. It includes several classic algorithms implemented from scratch in vanilla Python/NumPy, as well as some utilities. As a student, my goal is prioritizing clear, readable, and reproducible code over performance, so that more folks can better understand how AI works from a code-perspective.
Technologies: Python, NumPy, Matplotlib
For the final project of CS 556 (in grad school), my colleagues Jay, Jaydeep, and myself decided to explore a variety of supervised techniques (both parametric and otherwise) for identifying potentially hazardous objects detected close to our home planet. Along the way, we learned alot about outlier detection (which is typically thought to be an unsupervised problem). The dataset we used can be found on Kaggle.
Technologies: Python, Scikit-learn, NumPy, Matplotlib, Pandas, Seaborn, Jupyter Notebook
My expertise is in deep computer vision involving neural networks. You can trust me to provide you with the insight and implementation of the different algorithms applicable to your projects, all with a data-driven approach to problem-solving.
I will help craft breakthrough online experiences for your users. Allow me to translate your project requirements into scalable, maintainable, and highly-performant code using modern web frameworks.
I walk the line between theory and implementation, and will help you scope problems so that they not only become opportunities to invent new products, but also avenues to publish fundamental research.