Problem
Security and financial data produce more signals than a person can consistently evaluate without structured analysis and prioritization.
PythonScikit-learnPandasAPIsMachine LearningCase study / interactive demo
Case study
From problem to working system
My role
AI product engineer and full-stack developer
Solution
Suite of AI-driven agents for detecting cybersecurity threats and analyzing financial market behaviors using ML, predictive modeling, and real-time API data feeds with comprehensive visualization dashboards.
Architecture
The implementation combines the following technologies and system concerns.
PythonScikit-learnPandasAPIsMachine Learning
How it was built
- Built a suite of Python agents that pull live market and security telemetry through REST APIs and normalise it into a common feature frame for modelling.
- Implemented predictive models over financial time series with train/test separation by date to prevent look-ahead bias.
- Built visualisation dashboards that pair each prediction with the features that drove it, so an analyst can sanity-check the model rather than trust it blindly.
Security decisions
- Built a suite of Python agents that pull live market and security telemetry through REST APIs and normalise it into a common feature frame for modelling.
- Trained scikit-learn classifiers for threat detection on labelled network features, tracking precision and recall rather than accuracy because the classes are heavily imbalanced.
Major challenges
- Trained scikit-learn classifiers for threat detection on labelled network features, tracking precision and recall rather than accuracy because the classes are heavily imbalanced.
- Engineered pandas pipelines for cleaning, resampling, and feature derivation, with explicit handling for missing intervals instead of silent forward-fills.
- Structured the agents to fail safe on API errors and rate limits, degrading to the last known good snapshot instead of emitting fabricated values.
Verified evidence
Results and measurable impact
- Implemented predictive models over financial time series with train/test separation by date to prevent look-ahead bias.
- Engineered pandas pipelines for cleaning, resampling, and feature derivation, with explicit handling for missing intervals instead of silent forward-fills.
- Built visualisation dashboards that pair each prediction with the features that drove it, so an analyst can sanity-check the model rather than trust it blindly.
- Structured the agents to fail safe on API errors and rate limits, degrading to the last known good snapshot instead of emitting fabricated values.
No separate numeric outcome is documented, so this section shows shipped technical evidence without inventing metrics.
Screenshots and access
Product view
Interactive Demo
A scoped, fully functional recreation of this project's core feature runs below, live in your browser. Reset it, resize it, or expand it to full screen.
ai-cybersecurity.zainkhalilkhan.com
AI Agents
Specialist agent coordination
AI AgentsWorkspace4 updates
Agent LabAUTONOMOUS DECISION SYSTEMSlive
ticks processed
0
signals fired
0
Simulated agent reasoning
> Awaiting network telemetry.
Zain Khalil Khan
Client-side sandbox. State is in memory and nothing is sent to a server.