AI Security Models

Known Unknown

Cyber Security Rule based detections have certain limitations. Threats and signature based detections are required but most companies are augmenting these detections with ML to find unknown behavior and zero day malware/ ransomware.   CyberQ is proficient at helping customers design and implement custom ML models on several COTS platforms to deliver accurate and timely deep learning detection to your SOC.

Custom Models

Designing and creating custom models requires an expert level AI team that is focused on Cyber Security outcomes, that is the team at CyberQ.

DCJ

Protecting your digital Crown Jewel (DCJ) assets with custom AI models, or building out a robust Insider Threat program using AI is where many customers require custom models, on top of what a given SW vendor may already provide is a common requirement that CyberQ is focused on addressing.

Objectives

The business objectives are analyzed and a design for data ingestion, parsing and ML is designed with the customer to achieve the expected outcomes. Using ML to build unique detections with both grouping and chaining (multiple AI models) may be the result to achieve the specific customer objectives.

Training Models

Due to the nature of ML, models need to be trained over time or with a specific volume of data. Once these thresholds are achieved the CyberQ team will revisit a specific model to tune the model for accurate outcomes that can be leveraged by the SOC.

Training the Customer

Many times the concepts and implementation of ML is a new subject within the SOC team, or the insider threat teams. CyberQ works with each individual customer to help guide the teams with a customized learning path using the models and live data to help customers become self-sufficient over time.

Deliverables

CyberQ AI Custom models are delivered using a methodology to design usable outcomes, compliance driven, with training and a workflow that can be repeated over many models and data sets

AI and Analytics Test

Validation is a key step in the methodology that proves to the business that a given custom model is complete, and has the required outcome to make decisions. 

Engaging CyberQ

CyberQ AI Models program can be purchased as a complete program to design, implement, and rollout into production a mature customized ML solution using COTS SW that can be integrated into the main SOC workflow.  The program is focused on achieving results from the customers platform.

Quality Data Creates

Quality Results

The backbone of quality ML models is quality Data. CyberQ will work with the customer to audit and modify data ingestion to ensure quality outcomes from any Custom ML model.