Hi, thanks for visiting my page. My name is Salman. I'm a self taught software engineer and data scientist with a primary focus on computer vision problems.
Currently, I'm working at Stainless AI where we're pushing the limits of computer vision in Sports. My role varies from building neural networks for tasks such as image classification, object detection, semantic segmentation etc. to designing and developing deep learning software for both cloud and embedded systems. Most of my work here is generally based off of (but no limited to) TensorFlow and OpenCV and are written in C++ and Python.
In the past, I've worked at a financial company where I bootstrapped a data science department from scratch and developed a machine learning product that increased underwriting efficiency by 25%. The product used NLP techniques to automate a tedious, time consuming data collection process from customers' bank accounts and summarized relevant details in a manner that makes it easy to underwrite. The tech stack used was Python (Flask framework, scikit-learn for ML), Windows Server, and IIS.
Further, I've volunteered at XSCAPE Games—a video game startup that is working to build a real life mario cart game. At XSCAPE, I designed a controller system and a motion control system that allowed players to drive an RC using a simulator and finalized their prototype. I also served as an advisor for other developers.
In terms of education, I have a Bachelor's in Mathematics and machine learning certificate from Springboard bootcamp. For my capstone project, I've designed and image super resolution model and deployed it as a free public service.
If you're looking for a smart, talented, hard-working engineer to join your team and assist you in your projects, shoot me a message on LinkedIn or email. I'd be happy to see how I can help.
My capstone project for Springboard. For this project, I worked 1-on-1 with a mentor to build a model that increases the resolution of an image 2x using a residual neural network. The model implemenation is in tensorflow/keras. The deployment uses Flask as the web framework and a free Linux server from pythonanywhere.com
A startup project that I volunteered for. I designed the controller and motion control backend for the CXC simulator. The incoming telemetry data was trasmitted over UDP using a Pixracer mounted on the RC car. The prototype was designed using Python (primarily Pygame and pymavlink)
A project designed to demonstrate the applicability of machine learning tools to automate underwriting procedures. Experience from this project had helped tremendously in delivering my ML product at Wide Merchant Group
View ProjectA group project during my final quarter at UCLA. In this project, we taught an AI to produce handwritten numbers. The model implemenation was in Keras. We used gaussian mixture sampling to sample from the encoded vector space
View Project | View Paper