Jacob Bartel

Who am I?

I am currently a Senior Sofware Engineer at Google focused on speaker-identification machine learning models for both production and research efforts.

I have been at Google since 2014, and have worked in research, ML modeling, evaluation systems, and software infrastructre in a variety of areas including Social Networks, Content Recommendation, and Speech

I also dabble in multiple hobbies for fun including photography, cooking, and cycling.

Interests

My area of interest encompasses multiple different areas, mostly due to my involvement in different projects in my education, graduate work, and career. These include

I enjoy working on speaker-identification models using machine learning techniques.

I have a keen interest in generating predictions and recommendations for information sharing using a variety of data mining and machine learning techniques.

I also enjoy working with collaborative user interface generation. Specifically, I have worked with generating collaborative applications using the Java AWT, Swing, SWT, and GWT toolkits.

Education History

I received my Ph.D. in Computer Science from UNC-Chapel Hill in 2015 for work conducted under my advisor Prof. Prasun Dewan. My work focused on recommendations to assist users when sharing information at large scales

I received my M.S. in Computer Science in May 2012 from UNC-Chapel Hill

I received my B.S. in Computer Science with a minor in Mathematics from UNC-Chapel Hill in May 2009.

Dissertation

Predictions to Ease Users' Effort in Scalable Sharing

Document: PDF
Recorded Talk: YouTube
Slides with narration: PowerPoint; YouTube

Publications

Pai Zhu, Dhruuv Agarwal, Jacob W. Bartel, Kurt Partridge, Hyun Jin Park, and Quan Wang. Synth4Kws: Synthesized Speech for User Defined Keyword Spotting in Low Resource Environments. Interspeech SynData4GenAI 2024 Workshop. 2024.

Document: arxiv; Private

Hyun Jin Park, Dhruuv Agarwal, Neng Chen, Rentao Sun, Kurt Partridge, Justin Chen, Harry Zhang, Pai Zhu, Jacob Bartel, Kyle Kastner, Gary Wang, Andrew Rosenberg, Quan Wang. Adversarial training of Keyword Spotting to Minimize TTS Data Overfitting. Interspeech SynData4GenAI 2024 Workshop. 2024.

Document: arxiv; Private

Hyun Jin Park, Dhruuv Agarwal, Neng Chen, Rentao Sun, Kurt Partridge, Justin Chen, Harry Zhang, Pai Zhu, Jacob Bartel, Kyle Kastner, Gary Wang, Andrew Rosenberg, and Quan Wang. Utilizing TTS Synthesized Data for Efficient Development of Keyword Spotting Model. Interspeech SynData4GenAI 2024 Workshop. 2024.

Document: arxiv; Private

Di Huang, Jacob Bartel and John Palowitch. Recurrent Graph Neural Networks for Rumor Detection in Online Forums. KDD MIS2 2021.

Document: arxiv; Private

Jacob W. Bartel and Prasun Dewan. Towards Evolutionary Named Group Recommendations. Computer Supported Cooperative Work, Volume 27, Issue 3-6. December 2018.

Document: ACM Digital Library; Private

Danwei Li, Jacob Bartel, and Prasun Dewan. Extending the Gmail User-Interface to Leverage Prediction of Response Times and Hierarchical Recipients. Proceedings of 16th European Conference on Computer-Supported Cooperative Work-Panels, Posters and Demos. 2018.

Document: EUSSET; Private

Ziyou Wu, Isabella Huang, Xubin Zheng, Jacob Bartel, Andrew Vitkus, and Prasun Dewan. A test-bed for facebook friend-list recommendations. EICS '15: Proceedings of the 7th ACM SIGCHI Symposium on Engineering Interactive Computing Systems. June 2015.

Document: ACM Digital Library; Private

Connor Hamlet, Daniel Korn, Nikhil Prasad, Volodymyr Siedlecki, Eliezer Encarnacion, Jacob Bartel, and Prasun Dewan. User-Interfaces for Incremental Recipient and Response Time Predictions in Asynchronous Messaging. Proceedings of the 20th International Conference on Intelligent User Interfaces Companion (IUI Companion 2015). April 2015.

Document: ACM Digital Library; Private

Preeti Arunapuram, Jacob Bartel, and Prasun Dewan. Distribution, Correlation and Prediction of Response Times in Stack Overflow. 10th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing (IEEE CollaborateCom 2014). Oct 2014.

Document: Private

Andrew Ghobrial, Jacob Bartel, and Prasun Dewan. Towards Composable Prediction of Contact Groups. 10th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing (IEEE CollaborateCom 2014). Oct 2014.

Document: Private

Jacob Bartel and Prasun Dewan. Evolving Friend Lists in Social Networks. 7th ACM Conference on Recommender Systems (RecSys). October 2013. (Acceptance ratio: 0.34)

Document: ACM Digital Library; Private
Presentation: Poster

Jacob Bartel and Prasun Dewan. Towards Hierarchical Email Recipient Prediction. 8th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing (IEEE CollaborateCom 2012). Pittsburgh, PA. Oct 2012. (Acceptance Ratio: 0.25)

Document: IEEE Explore; Private
Talk: PPTX; YouTube

Jacob Bartel and Prasun Dewan. Towards Multi-Domain Collaborative Toolkits. Proceedings of the ACM 2012 conference on Computer Supported Cooperative Work (ACM CSCW 2012). Feb 2012. (Two-round review process)

Document: ACM Digital Library; Private
Source code: GitHub

Previous Work

I was also a TA in Spring and Fall of 2014 working with Dr. Jay Aikat for Comp 110 - Introduction to Programming. As a part of my TA duties, I developed and managed the backend of a automated grader program developed originally by Prof. Prasun Dewan and Joshua Monson. This tool automatically grades assignments and allows multiple and distributed TAs the ability to quickly grade and review assignments from over 300 students. The Grader tool is available at this Github repo

I have helped develop a virtual toolkit for transparently creating a collaborative GUIs using the Java AWT, Swing, SWT or GWT toolkits. The specific toolkit can be chosen at runtime.
The source code for the project can be found here.

I worked as an intern programmer at the Smithsonian Institution Archives in the Summer of 2008. One of my major code projects was developing a script to combine the outputs of the Jhove and DROID scanning tools into one easily understandable output. The project is detailed here.

Who am I?

Interests

Education History

Dissertation

Publications

Previous Work

Related presentations