Correlates multiple features (Threat signatures, total connections, total Bytes sent/received etc.)

Analyzes the cumulative distribution function of features to detect outliers in the data.

Surfaces the alerts on the UI. (high/medium/low severity)

Created a context-based anomaly detection feature (e.g. a computer running IOS)

Used the idea of a research paper which uses likelihood based approach to create probability distributions of categorical data.

Gives a measure of negative correlation of categorical features based on probability of co-occurrence.

Built a time-based anomaly feature to detect activity/inactivity of a device at an unusual time.

Uses the probability of activity in the past detected from IOT sensors to detect unusual activity/inactivity.

Experience in:

Extraction (S3, Elasticsearch, MongoDB), Wrangling (Spark/Pandas operations), Visualizing (Matplotlib)

Deploying features on staging/production (AWS instances)

Creating streaming and batch jobs using shell scripting and scheduling them on Azkaban.