Technical feasibility
Review the system requirements, available data and integration constraints. Identify the work needed to test the proposed approach.
Expertise
Cyber & ML
engineering
Detection, threat intelligence, secure automation and bespoke datasets for defence and unmanned systems.
SensingCyberSoftwareData
Sensor systems
Long-range detection
& sensor fusion
Detection and fusion algorithms for systems that need to interpret several sensor feeds.
We combine sensor inputs while accounting for timing, alignment and uncertainty. The work covers detection models, fusion logic and the interfaces that carry results into a wider system.
- Applications
- Long-range detection and situational awareness.
- Inputs
- Multiple sensor types, asynchronous observations and incomplete data.
- Evaluation
- Detection performance, false-alarm rates and sensitivity to data gaps.
Cyber security
Threat intelligence
& response
We analyse security logs and threat data to help teams detect suspicious activity and investigate incidents.
We develop methods for event classification and threat analysis, with attention to false alarms and changing activity patterns. Our team’s LLM4Sec research benchmarks language models for security-log analysis.
- Work
- Security-log analysis, event classification and threat assessment.
- Evaluation
- False alarms, changing activity patterns and the quality of source data.
Software & systems
Secure automation
& verification
We verify software behaviour and design controls for automated systems.
Access controls, interface definitions and failure handling are part of the engineering work. Tests connect software behaviour to system requirements, including what happens when inputs are missing or components fail.
- Work
- Software verification, secure interfaces and controlled automation.
- Controls
- Permissions, audit trails and change management.
- Verification
- Requirements-based tests and failure-case analysis.
Data & machine learning
Data for
unmanned systems
We design, collect and analyse datasets for sensing and unmanned systems.
Collection and labelling follow the sensing task and its operating conditions. We assess coverage, identify gaps and evaluate models on data kept separate from training.
- Work
- Collection planning, annotation, dataset analysis and model evaluation.
- Data quality
- Provenance, label accuracy and coverage of operating conditions.
Working together
Technical
partnerships
We take on specific development tasks and contribute cyber and machine-learning expertise to larger defence programmes.
Software & ML
Develop datasets, algorithms and software components for the application, with tests that track their performance.
System verification
Connect components through documented interfaces and verify their behaviour against system requirements.
Project enquiries
Contact Vaktros
Tell us what you’re building and where you need specialist cyber or machine-learning expertise.
Contact our team