Write user needs and system requirements for the CGM product

Write verification tests for the system requirements and execute tests. Utilize tools like a DMM and SQL to verify signal measurements and data records respectively

Test GraphWear app product for functionality bugs and submit defect reports to Software team

Test system-level integration (apps, CGM device, cloud architecture) and submit defect reports

Document system use-cases and identify failure-modes for testing

Write Work Instructions for process tasks like Device Management in clinical trials

Support Data team with data analysis efforts and maintain the interactive dashboard of signal performance metrics

Use the DSP (digital signal processing) algorithm stack to convert voltage signals to calibrated plasma glucose concentrations

Use Git-based tools SourceTree and BitBucket to add calculation and visualization tools, as well as signal correction algorithms to master branch

Create an interactive dashboard with Plotly Dash for executive team to view real-time distribution analysis of data

Analyze key metrics that describe signal quality and data compliance to FDA iCGM standards by generating 2D and 3D data visualizations in Plotly by various category groupings

Understand database architecture and collaborate with Data Engineers to address data requirements (e.g. type, format) for effective data analysis. Ensure database adjustments are stable across the system (app, cloud, devices)

Cleaned and standardized scraped data to optimize inputs to Recommendation Engine in CE App

Compared NLP (Natural Language Processing) models TF-IDF, Word2vec, and BERT for Recommendation Engine in CE App

Used MLflow to track NLP model changes and assess model performance

Completed thesis on novel gut-brain PAC (phase amplitude coupling) to understand the association of visceral states and cognitive performance

Analyzed EEG (electroencephalogram) and EGG (electrogastrogram) data using Python data science & visualization libraries NumPy, SciPy, Pandas, Scikit-learn, and matplotlib to identify periods of inter-signal modulation

Processed timeseries data using filtering, Hilbert transform, and spectral analyses to isolate low-frequency gastric signals and high-frequency EEG signals

Led weekly meetings between project collaborators