Automating the detection of cyber attack campaigns via ML algorithms-Developed the prototype and proof-of-concept of unsupervised models to capture malicious activity from snort logs-Combined ML and graph algorithms to detect attack campaigns-Labeled snort logs into TTP classes of MITRE framework using NLP algorithms

ML R&D on fall risk prediction of dementia patients given irregularly-spaced temporal gait data and baseline clinical scores. Experiments involve design an implementation of ML models tailored to handle irregularly-spaced sequential data.

Introduction to Machine Learning (CSC311)

Algorithm Design and Analysis (CSC373)

Employed semi-supervised methods to predict severity of hydronephrosis from renal ultrasounds of infants. Model involved non-linear manifold learning from unlabelled data to reduce dimension for the downstream classification task.

ML R&D for multi-label document classification of medical notes from the Tuberculosis clinic. Employed Transfer Learning and NLP techniques on ML models to compensate for lack of sufficient data

Software Engineering at Input/Output(IO) Subsystem team of Intel Programmable Solutions Group.•Developed IO subsystem of Quartus software regarding my IP to meet FPGA chip hardware constraints and customer designs specifications•Collaborated with various teams to ensure code compatibility with other subsystems during development •Added software modelling of IO buffers of the new family of FPGA devices to Quartus•Identified unknown dependencies in the source code and documented correct procedures to modify the corresponding sections of the code in order to avoid software break down during future developments•Implemented regression tests to ensure code functionality during development•Addressed customers’ requests relating to my IP on a regular basis, by investigating their use case and debugging potential errors in the source code

Web developer working on TELS (Teamwork Effectiveness Learning System) online system used for educational purposes•Communicated with stakeholders (educators) to understand features required for the online system and designed a model accordingly•Developed new analytical features which aggregated inputs from student teams and provided assessment on the data•Designed a tracking system to efficiently query the database, summarize data and generate reports•Surveyed potential users on the usability and functionality of the new features and improved design through subsequent modifications •Software currently being commercialized and sold to other Universities.