Sajeeva Salgadoe
I am a data scientist with a strong security background and 8+ years of industry experience in data science, data security and wireless technologies.
I co-founded a successful service provider company (Routcom Inc) to provide broadband wireless service to residences in north of Toronto, ON.
I enjoy being challenged and my eagerness to cultivate knowledge led me to enroll in graduate studies. I am currently finalizing my Ph.D in computer science at UOIT. My research area is anomaly detection in low powered networks using Machine Learning.
My professional achievements include two highly regarded security credentials (CCIE (security) and CISSP) and number of conference & journal publications.
To summarize, I am an entrepreneur, a hard working individual with a strong data science and security background and more importantly; I am a husband and a proud father of four kids (yay !!!)
Selected Work [./]
Real Estate Value Estimator
Machine Learning based regression model parameterized by several critical attributes is developed to approximate the real estate property value. The prediction model can be queried by real estate owners to estimate their property values.
Service Availability Predictor
A Machine Learning model to predict wireless service availability of a particular address based on the location, existing customers in neighbourhood and other RF characteristics including radio signal strength and noise.
Traffic Usage Calculator
Dynamic traffic usage web portal provides near-real time data usage for wireless users. The portal provides real time detailed traffic information between source and the destination.
RF Property Monitor
RF properties including SSID, SNR, LQI, Channel Noise and the Link Distance are analyzed in near-real-time to monitor fluctuations in radio links. Historical data is used to build a prediction model to analyze the variation of expected and real values of RF parameters.
News Aggregator
News Aggregation application is able to collect all news items associated with a given web url. The model is based on scraping html content instead of relying on RSS feeds and other techniques such as RestFul APIs.
Emergency Alerts Generator
Emergency Alerts Generator collects alerts related to various categories including weather, political, epidemic, riots and terrorism, using different resources including RSS Feeds, Web Feeds and other APIs. The objective of the model is to develop a prediction model to determine the category and the severity of a particular alert.
Contact Extractor
Contact details related to embassies, consulates and other foriegn representatives are extracted from various sources. Application is configured to crawl a list of urls at regular intervals to collect most upto date information from various resources including html, text and PDF.
EHR Extractor
In this work, health data generated by a legacy cardiovascular apparatus are extracted to build a model to predict certain health anomalies of a patient.
Data Collection Projects
Details about several data collection projects related to my work as well as related to academic research can be found here. Some interesting work includes real-estate data scraper, auto inventory data scraper, news articles scraper and RF data collector using passive interception of wireless communication can be found.
Experimental Work
A several experimental work related to machine learning, deep learning, statistical analysis, as well as various libraries/frameworks including Scikit-Learn, Pytorch, TensorFlow, Spark, Django, Restful API, D3 JavaScript and WordPress can be found here.