NetSafe AI Edge Load Balancer
Inspiration
DDoS attacks cost the global economy tens of billions of dollars every year, pushing security-related downtime losses past $400B. Most of this damage happens because defenses react only after backend systems are already overwhelmed.
We wanted to build something proactive a system that stops malicious traffic before it causes downtime, lost revenue, and loss of user trust.
What it does
NetSafe is an intelligent edge load balancer that blocks malicious traffic at the edge and routes legitimate users in real time.
Instead of treating all traffic equally, NetSafe identifies attack traffic instantly and prevents it from ever reaching backend servers.
How we built it
NetSafe runs on Vercel’s Edge Network, allowing traffic to be analyzed as close to the user as possible.
We use a lightweight machine learning model trained on 175,000+ data points to detect malicious behavior. Legitimate traffic is routed to backend servers hosted on Render, while malicious requests are blocked.
A React-based dashboard provides real-time visibility into traffic, attacks, and server health.
Challenges we ran into
Our biggest challenge was making machine-learning detection fast enough to run at the edge without adding noticeable latency.
We also had to ensure reliable routing under attack conditions while keeping the system lightweight and responsive.
Accomplishments we’re proud of
We successfully combined edge computing, load balancing, and real-time bot detection into a single working system.
NetSafe is able to stop attacks in milliseconds before they impact backend infrastructure.
We also built a full end-to-end demo with live monitoring and traffic visualization.
What we learned
Stopping attacks early is far more effective than reacting at the server level.
Edge computing enables security decisions to happen faster and more efficiently.
We also gained valuable experience designing systems that remain reliable under adversarial conditions.
What’s next
Next, we plan to improve the detection model with richer behavioral signals and adaptive learning.
We want to add smarter routing based on latency, cost, and geographic location.
Long-term, NetSafe could evolve into a shared threat intelligence platform, where detected threats help protect all connected services.
Log in or sign up for Devpost to join the conversation.