Other Projects
Hackathon Highlights: Diet Vision Wins 3rd Place! 🏆
We’re thrilled to announce our victory at the recent hackathon, where Diet Vision claimed 3rd place! Diet Vision is our groundbreaking tool that redefines calorie tracking through automated logging with Continuous Glucose Monitors (CGMs), providing new ways to combat obesity and enhance health management.
What Sets Diet Vision Apart:
- Seamless Connectivity: Bluetooth LE syncing with CGMs
- Real-time Recognition: Computer vision to instantly identify foods
- Automated Tracking: Instant nutrient and calorie tracking, no manual entry needed
Technology Stack:
- Hardware: Raspberry Pi Zero prototype paired with Libre 2 CGM
- Backend: Flask server and Nvidia Triton Inference Server for scalable AI
- ML Pipeline: ONNX models for streamlined interoperability
- Computer Vision: Powered by OpenCV, TensorFlow, and Keras for precision image processing
- Embedded Systems: MicroPython on a custom PCB (upcoming)
Roadmap and Vision:
- Miniaturization: Custom PCB for a more compact device
- Enhanced Accuracy: Advanced ML models for improved food recognition
- B2B Opportunities: Integration with major CGMs, e.g., Abbott Diabetes Care Inc.
- Privacy First: Future implementation of federated learning for secure model updates
Market Potential:
- Targeting a $70 billion market, beginning with the 422 million individuals living with diabetes
- B2C launch strategy with B2B partnerships in the pipeline
Next Steps: FDA approval, a crowdfunding launch, and rapid scaling to bring Diet Vision to more users.
Meet the Team Behind the Innovation:
- Ernest Choi
- Fareed Khan
- Rafiea Ashraf
Check out our press release here and share your thoughts—what feature would you like to see in smart calorie-tracking?
https://www.youtube.com/watch?v=us5uuHK-qLk
Veterans Affairs MIT Hacking Medicine Hackathon: Team AIHR Takes 3rd Place!
We had the honor of participating in the first-ever Veterans Affairs MIT Hacking Medicine Hackathon, where our team, AIHR, built a groundbreaking AI tool that accelerates the hiring process by more than three weeks. In just 48 hours, we developed a working proof of concept that has the potential to make a meaningful impact in the hiring process at the VA.
About Our Solution: Our AI-driven solution is designed to optimize the hiring timeline, potentially saving the VA a month in processing time, leading to faster placements and enhanced efficiency in critical hiring practices.
Team Members:
- Nikhil Sharma
- Nikola Andric
- A. Mana Birgani
- Hillary Umeh Jr.
- Jose Urdaneta
Special Thanks: A heartfelt thank you to Indra Sandal, PhD, MBA, and Veronica Herst for their invaluable insights into the VA’s processes, which were instrumental in shaping our approach. Additionally, we appreciate the guidance of over two dozen mentors who supported us, especially Ryan Holt and the Microsoft help desk team, who dedicated countless hours helping us with Microsoft Azure troubleshooting.
Stay tuned as we further develop this promising project in the upcoming Makerthon!
GrandHack Success: Diapatch Wins 3rd Place! 🏅
A Design-Centric Approach:
Unlike traditional hackathons, GrandHack emphasized a design-first mindset, challenging us to dive deep into understanding our stakeholders. We asked:
- Who are our end-users?
- What specific challenges do they encounter?
- How can Diapatch be tailored to meet their unique needs?
Reflections on the Experience:
This hackathon was a powerful reminder of the importance of thoughtful design and the value of collaborating with a diverse, multidisciplinary team. I’m more motivated than ever to approach future challenges with renewed creativity and empathy for users.
Looking forward to the next big idea!