Software Engineer
Digital
Portfolio
I'm a software engineer who loves diving into new projects and building thoughtful, user-centered experiences (especially ones involving emerging technologies).
Right now, I’m working full-time on my own project—Sort My Shoebox—but I’m also open to the right opportunity in engineering or product management.
Keep scrolling to see what I’ve been building, or feel free to reach out at grace@graceraper.com.
Ghost Pacer is AR-Fitness startup that produced custom augmented reality glasses for outdoor applications. While I was there, we grew to seven figures in annual revenue & received funding from Techstars before being bought by Peloton in 2024.
I served as both COO & Lead Software Engineer. My primary focus was on designing and engineering a lightweight AR fitness experience that combined multi-sensor data (GPS, IMU, biometrics) to create a responsive virtual running partner that adapts to the user's environment and performance.
This software was highly optimized to run on lightweight hardware while processing complex spatial calculations and managing connectivity with multiple Bluetooth peripherals (including watches, heart rate monitors, and mobile devices).
Check out the website to learn more!
SortMyShoebox is an AI-powered photo management platform that makes it simple to scan and digitally organize old family photos. Using computer vision and machine learning, it transforms the daunting task of digitizing a print photo collection from a project that takes months to a project that can be completed in minutes.
Modern photo management platforms rely on metadata (i.e. dates, locations, events) in order to organize images. When photos are digitized, this metadata is absent, making the resulting digital collections difficult to navigate in modern systems. SortMyShoebox's metadata prediction engine uses facial recognition, age estimation, family history, and more in order to predict when an image was taken - often down to the exact date.
I've designed and built the entire platform from the ground up. The software automatically tags family members, estimates photo dates, groups related images, and generates meaningful captions to preserve the stories behind each moment.
Ready to rediscover your family memories? Visit the SortMyShoebox website to get started.
I'm currently prototyping Remme as an extension/pivot of SortMyShoebox.
Remme uses the metadata-predictions engine behind SortMyShoebox to revolutionize elderly care through AI-powered reminiscence therapy. By transforming digitized family photos into interactive therapy sessions, we're creating meaningful connections for seniors experiencing cognitive decline or isolation.
Studies have shown that reminiscence therapy can significantly improve seniors' quality of life - reducing depression, enhancing cognitive function, and decreasing reliance on antidepressants. Remme makes this powerful therapy accessible 24/7 through an intuitive, FaceTime-like interface.
Our AI companion acts as a gentle guide through cherished memories, dynamically adapting conversations based on engagement and emotional responses. This creates a safe, judgment-free space where seniors can revisit their life stories at their own pace, maintaining dignity and independence.
See how Remme will transform elder care at tryremme.com.
I'm retrofitting a vintage rotary dial telephone into a modern jukebox and Bluetooth headset. Users can add songs to their Apple Music/Spotify queue by dialing a 3-digit code from a songbook. The phone also functions as a Bluetooth headset, allowing users to make and receive calls from their smartphone through the classic rotary handset.
Coming in July: follow my step-by-step build guide on Medium to create your own.
CaspirCam is an iOS app I built that uses AR technology to empowers users to improve their photos. By selecting a template pose or uploading an inspiration photo, users can view a virtual guide to replicate the pose in real-time. The app provides audio and visual feedback to help coach users as they adjust their posture and facial expression, ensuring they look their best. CoreML and the iPhone’s camera work together to analyze body alignment and offer precise step-by-step guidance.
View the product website to learn more!
Created a web-based visualization frameworks that enables users to explore the relationship between historical redlining maps and current-day census data.
Some sources of inspiration for this project include the New York Times "How Decades of Racist Housing Policy Left Neighborhoods Sweltering" Article published in 2020 by Brad Plumer and Nadja Popovich as well as the "Mapping Inequalities Project: Redlining in New Deal America" out of the Digital Scholarship Lab at the University of Virginia.
Check it out: Mapping The Lasting Effects of Redlining
This interactive 3D portfolio was built using Three.js and WebGL, with 3D models created in Blender. The structure was influenced by bokoko's award-winning portfolio and pandawoan's tutorials on creating interactive 3D websites.