Inspiration
We were inspired by the ever-growing complexity of financial fraud and the need for better compliance tools in the industry. Traditional systems often struggle to keep up with digital transactions, so we wanted to create something smarter—something that combines blockchain’s security with AI’s ability to detect fraud in real-time. Our goal was simple: build a system that can help financial institutions stay ahead of fraud while ensuring compliance.
What it does
Our API offers financial institutions two key features:
Transaction Logging (Blockchain): It records every transaction on a secure, immutable blockchain. This creates a transparent, tamper-proof record that can be easily accessed for audits and compliance checks.
Fraud Detection (AI-powered): The API scans each transaction in real time using an AI-driven model (autoencoder). If the transaction appears suspicious, it’s flagged, and alerts are sent to a compliance dashboard for further investigation.
The API seamlessly integrates with existing transaction channels, ensuring financial institutions can prevent fraud while staying compliant without disrupting their workflow.
How we built it
Blockchain Development: We created a custom blockchain and used Flask for routing and Python for the core logic. Smart contracts were deployed to validate and enforce business rules on the blockchain.
AI-powered Fraud Detection: We used TensorFlow to train an autoencoder model to detect fraud. It learns the "normal" patterns and flags anything that deviates.
Frontend Interface: We built a React.js dashboard to let compliance officers easily interact with flagged transactions and blockchain records.
DevOps & Scalability: Docker made our deployment process smooth, and Git/GitHub helped the team stay in sync as we built the system.
Challenges we ran into
Integrating Blockchain and AI: Getting blockchain and AI to work seamlessly together was tricky, but we nailed it by carefully connecting real-time fraud alerts to blockchain logs.
Building a Scalable Blockchain: We needed our custom blockchain to scale efficiently with transaction volume. Optimizing it for performance was a big hurdle, but we made it happen.
Fine-tuning AI Detection: Training the AI model with diverse transaction data was a challenge, but once it was set up, the fraud detection system worked like a charm.
Accomplishments that we're proud of
What we learned
Blockchain's Strength in Trust: Blockchain's immutability ensures transactions are secure and auditable, giving institutions an unbreakable proof of compliance.
AI's Power in Fraud Detection: AI—specifically, an autoencoder model—can quickly spot fraud by learning transaction patterns and flagging anomalies in real-time.
The Magic of Combining Both: Blockchain + AI = powerful risk management. The blockchain logs transactions while AI scans for fraud, creating a highly reliable and scalable solution for financial institutions.
What's next for Sentry
This project was an exciting challenge. By combining blockchain's transparency with AI’s fraud detection, we created a powerful tool that helps financial institutions manage risk, prevent fraud, and ensure compliance. The road wasn’t always easy, but seeing the system come together—helping prevent fraud in real time—made it all worth it. We’re proud of what we built and excited about the impact it can have on the financial world.
Log in or sign up for Devpost to join the conversation.