Meet the Expert: Upstream Detectors Explained

Shachar Azriel, VP Data, explains Upstream’s unique approach to building and formulating detectors on the Upstream Platform. Drilling down to what detectors are exactly, who can build them, and why it’s so important to empower customers to build their own detectors on the platform for tailored cybersecurity detection and response.

 


 

Script

What’s a detector? 

A detector is a piece of logic. In practice, this means it’s pseudo code, a basic or advanced one that answers a defined use case. 

Anything in the world has a data signature – When we build these detectors, we need the ability to refine what these signatures are, and where to find them. Our expertise is translating a cyber attack scenario to a simple logic and understanding where we can find it in the data we receive from the OEMs.

Upstream aims to democratize detector creation – Based on our experience, some of the best detectors are built by our customers. The customer has the most valuable insights into their fleet features and the specific use they want to deal with.  The customer is sometimes part of the design process or has access to threat models that provide them with an understanding that we don’t have. 

Upstream helps our customers to translate this unique knowledge into signatures, then analyzes and validates them based on their data. 

How do we know what to look for? 

Based on the fact that we monitor more than 12 million vehicles around the world and collect automotive threat intelligence with AutoThreat we can determine the most common and severe attack vectors and threats in our ecosystem.

Based on that knowledge, we use our expertise to translate these threats to data signatures and extract them from the OEM’s data. 

For example, when a hacker tries to take over a vehicle via an API, there are several stages you need to be able to classify and detect, and if you don’t have the expertise, this kind of attack will stay at an abstract level that can’t be detected or lead to a detector that will generate a high false positives alerts.

Who can build detectors? Does it require any know-how or training?

Upstream’s platform is built and designed to enable customers to build their own use cases, therefore it’s user-friendly and intuitive. With the basic training we provide our customers and partners, almost anyone can build their own detectors!

The user needs to understand which threat he wants to mitigate and which data is collected from their vehicles and servers. From that point he can use our platform to build any kind of pseudo code he desires and validate it.

Newsletter Icon

Subscribe
to our newsletter

Stay up-to-date on the latest trends, emerging risks, and updates

Contextual Cybersecurity: Digital Twins Power AI-Driven Threat Detection Across the Connected Vehicle Ecosystem

The automotive industry is undergoing a profound transformation, with software-defined vehicles (SDVs) at the forefront of this revolution. Vehicles now generate unparalleled volumes of data,…

More Details

Upstream’s 2025 Global Automotive Cybersecurity Report Executive Summary

The Automotive and Smart Mobility ecosystem experienced a sharp increase in cyber threats throughout 2024, with large- scale ransomware attacks causing unprecedented disruption. As cyber…

More Details

2025 Predictions: The Future of Automotive Cybersecurity

As cyber threats grow in scale and complexity, industry leaders weigh in on the challenges and innovations that will define 2025. From AI-driven defenses to…

More Details

The Hidden Cybersecurity Risks of the EV Charging Ecosystem – A Critical Look

As electric vehicles become a staple of sustainable mobility, the infrastructure supporting them—particularly EV charging stations—is growing rapidly to meet demand.

More Details

Proactive Quality Detection Solution Overview

Accelerate the resolution of vehicle quality issues by 30% with ML and AI-based early detection and root cause investigations.

More Details

Proactive Quality Detection

Accelerate resolution of vehicle quality issues with early detection and root cause investigations.

More Details
Skip to content