Inspiration To generate realistic synthetic data for AI model training without compromising privacy.
What it does Creates anonymized, structure-preserving synthetic datasets from real-world data.
How we built it Used FastAPI for the backend, Python libraries for data modeling, and a Next.js dashboard for interaction.
Challenges we ran into Balancing data utility with privacy while avoiding overfitting to original data.
Accomplishments that we're proud of Successfully generated synthetic datasets that mimic statistical patterns of real data.
What we learned Techniques in differential privacy, generative modeling, and data validation.
What's next for Data Synthesizer Add support for time-series and image data generation with export to major ML pipelines.
Built With
- access
- auth:
- control
- database:
- frontend:-next.js-14-backend:-fastapi-cloud:-aws-(ec2
- postgresql
- role-based
- s3)
Log in or sign up for Devpost to join the conversation.