Kavi Sachania | Projects Experience About |
Software Engineer
|
I am a software engineer with a master's and bachelor's degree in computer science after three and a half years at Santa Clara University. My current work at Performance Star, a data science R&D lab, exists at the intersection of software engineering and data science consulting. |
Forward Pedestrian Location IndicatorIn order to help improve driver awareness, I built an object detection and lighting system that indentifies and communicates in real time the locations of pedetrians on the road using a camera and LED light strip. A Jetson Nano with a GPU performs inference on camera video using a YOLO convolutional neural network while an Arduino drives the LED light strip that provides the driver with the locations of pedestrians at all times. Built with Python, C++, TensorRT, and OpenCV |
Fantasy Football Draft Strategy SimulatorI built draftstrategysim.com (no longer live) so that fantasy football fans like myself can experiment with different fantasy draft strategies and receive an accurate projection of the team's weekly scoring. The web application built with Python Django, PostgreSQL, and React runs a Monte Carlo simulation with 100,000 iterations on a lineup that the user selects and displays the chance of a selection starting in any given week during the season. Weighing player scoring by their start percentage creates a better representation of team value than widely-used arbitrary methods. Built with Python Django, React, MongoDB, and AWS |
College Football Computer RankingsCreated my own college football computer rankings using Python, College Football API, and AWS Sagemaker. Multi Order Win Ratings and methodology can be viewed at mowratings.com (no longer live). Built with Python Django, React, MongoDB, and AWS. |
Stock Pattern Email Alert SystemIn order to automate my stock discovery process, I built a Java Spring application that finds stocks that have reached their price support line and sends customizable weekly email alerts of all stocks that match certain price pattern criteria. Email alerts can be configured using a web application built using React and the REST endpoints on the Java Spring application with PostgreSQL database. Stock prices are fetched from Finnhub's API, and email messaging is handled by AWS Simple Email Service. Demo login email: demo@tradealert.money Demo login password: Password123 Built with Java Spring, React, PostgreSQL, and AWS |
Full Stack Software Engineer @ HavaloJune 2017 — September 2019
Technologies: Java, JavaScript, MySQL, HTML, CSS, Git, Docker, AWS, Maven |