Research Specialist III

San Diego State University Research Foundation•Calexico, CA

6d•$5,859•Onsite

About The Position

The Research Specialist III position is an advanced-level role requiring specialized knowledge, skill, and experience relevant to the project. The incumbent will work independently to prepare detailed plans for a research study focused on testing the performance of commercially available vape detectors and exploring the potential of converting consumer air monitors into vape detectors using machine learning. This two-year project involves laboratory experiments measuring indoor emissions and developing machine learning models to identify specific indoor sources. The role includes supervising and directing a small work group of students and research assistants.

Requirements

Equivalent to graduation from a four-year university.
Three years of experience in technical research or statistical work experience.
Experienced with indoor air quality monitoring, using research (e.g., SidePak, OPS) and low-cost air sensors (e.g., PurpleAir, Atmotube).
Familiar with Python and MATLAB programming languages.
Familiar with machine learning model training and inference procedures (e.g., logistic regression, artificial neural network).
Must reside in California and live within a commutable distance from SDSU at time of hire.

Nice To Haves

Experienced with field measurements of secondhand smoke and vape.
Familiar with Logistic and Artificial Neural Network modeling.
Published scientific journal(s) to support the qualifications above.
Professional Engineer (PE) license in Environmental Engineering.

Responsibilities

Take the lead on laboratory experiments testing air sensor performances (sensitivity and specificity) in identifying vaping, smoking, and other indoor emissions.
Perform machine learning modeling, including intensive hyperparameter tuning for neural network models.
Co-instruct two graduate students for emission testing.
Assist the Principal Investigator (PI) with the preparation of journal publications and annual reports for the two-year project.
Purchase requisition for commercial vape detectors and consumer air monitors.
Purchase requisition for vaping, smoking, and other indoor source materials.
Prepare and organize instruments in the experimental room.
Perform indoor emission tests for different indoor sources (e.g., vaping, smoking, cooking, dusting).
Evaluate performance of commercial vape detectors (sensitivity and specificity).
Assess potential of consumer air monitors in identifying vaping (sensitivity and specificity).
Ensure personnel safety during emission testing (e.g., N95 mask, exposure alert system).
Ensure a rigorous scientific protocol for experimental procedures.
Discuss and update experimental results with PI.
Compile and synchronize real-time measurements of air monitors in each experiment.
Label air monitoring data based on emission sources (e.g., vaping, smoking).
Aggregate data across different emission experiments based on source types.
Transform measurement profiles into normalized features for machine learning.
Construct training, cross-validation, and test data sets for each air monitor.
Perform hyperparameter tuning for neural network modeling (e.g., systematically testing different numbers of hidden layer/units, activation functions, regularization parameters, learning rates, batch sizes, numbers of iterations).
Train optimal neural network source identification models for each air monitor.
Test model classification performance (accuracy, precision, recall, F-score).
Compare optimal neural net models with other machine learning models.
Upload developed ML algorithms to mini PC (e.g., Raspberry Pi) for real-time source alerts.
Assist PI with journal publications on the performance of commercial vape detectors and the newly developed low-cost vape alert system.
Assist PI with annual reports for the two-year project.
Co-instruct two graduate students for experimental work.
Co-instruct two graduate students for labeling the source-specific data.