• Tach Natural Language Processing, Machine Learning and Deep Learning courses to doctoral students.
• Advise doctoral students.
• Research in Arabic NLP.
• Led a team to develop the largest Arabic NER corpus (560K words).
• Provide recommendation to the doctoral program (responsibility as a committee member).
• Organizer of the Arabic NER shared task at Arabic Natural Language Processing Conference (ArabicNLP), co-located with EMNLP 2023, responsibilities include developing baseline models, preparing Python notebooks and data, author and review publications.

• Developed and led Synthetic Panels powered by custom LLMs to generate synthetic responses to surveys.
• Led and supervised Qualtrics LLM fine-tuning across multiple projects including Insight Explorer and EX comment summary.
• Led and developed Qualtrics anonymization model.
• Mentored multiple junior scientists.

• Developed and led Synthetic Panels powered by custom LLMs to generate synthetic responses to surveys.
• Led and supervised Qualtrics LLM fine-tuning across multiple projects including Insight Explorer and EX comment summary.
• Led and developed Qualtrics anonymization model.
• Mentored multiple junior scientists.

• Developed and led Synthetic Panels powered by custom LLMs to generate synthetic responses to surveys.
• Led and supervised Qualtrics LLM fine-tuning across multiple projects including Insight Explorer and EX comment summary.
• Led and developed Qualtrics anonymization model.
• Mentored multiple junior scientists.

• Audio data preprocessing using short-time Fourier transform (STFT).
• Seq2Seq modeling for applications at the interaction of speech and computer vision.
• Generating speaker embeddings.
• Voice synthesis (vocoder)
• Generative models including GANs and VAEs.
• Named entity recognition: seq2seq modeling to built NER models based on BERT and LSTM architectures.
• Temporality extraction: this is an extension to NER which allows the ex-traction of temporal events.
• Language expansion: extending NER English model to other languages.
• Data processing, active learning and ETL: data pre-processing, tokenization, active learning and database.
• Mentoring: mentoring interns, new team members and leading projects.

The goal of the ”Advanced Development of an Open-source Platform for Web-based Integrative Digital Image Analysis in Cancer” is to develop open-source integrative technologies to facilitate analysis of data provided by the National Cancer Institute’s The Cancer Genome Atlas (TCGA). Some of my responsibilities include:

• Developing annotation capabilities that allow users to markup image data,
• Assisting in transforming the current Cancer Digital Slide Archive to an open-source software,
• Providing computational image analysis capabilities,
• Developing integrative analysis dashboard.

• Computational health,
• Patient similarity, similarity learning, graph based similarity and treatment recommendation,
• Predictive modeling,
• Machine learning including clustering, classification and prediction.

Besides research, I am very hands-on, I:

• Configured cloud computing clusters on Amazon Web Services (AWS),
• Installed packages on AWS clusters including Apache Spark and databases,
• Developed web services running on an AWS instance,
• Designed web services based on Health Level 7 Fast Health Interoperability Resources,
• Managed servers and data access.

Additional responsibilities included:

• Working with graduate and undergraduate students on various projects and applications,
• Mentoring graduate and undergraduate students,
• Assigning tasks to the students,
• Publishing scientific papers,
• Teaching.

This work was done in collaboration with governmental agencies and companies including:

• Center of Disease Control (CDC),
• Union Chimique Belge (UCB), a pharma company headquartered in Brussels, Belgium,
• Children’s Healthcare of Atlanta.

Leveraging Information Technology for Hi-Tech and Hi-Touch Care (http://light2.missouri.edu/) is a $13.3 million grant, federal government funded aimed at improving the health and lowering the cost of a population of 10,000 patients. As an analyst, my responsibilities included:

• Performing patient risk stratification using rules based and cluster analysis,
• Designing and implementing a web based portal for tracking how patients attribution to physicians,
• Collaborating with Cerner to design a datamart to support meaningful use and data analytics,
• Performing Extract, Transform, Load (ETL) processes from EHR to the datamart,
• Developing scripts that implements some of the business logic (deployed later by Cerner).