Embedded Files
I am a computer scientist working on graph algorithms and machine learning.
About Me
About Me
I am a Senior Software Engineer at Google Research, leading its Contextual Sampling team, reporting to Vahab Mirrokni. My core work involves developing highly scalable graph algorithms and techniques, crucial for optimizing products like Google Search, YouTube, AdWords, Play, and Maps. A significant part of my recent focus has been on data curation for Google's Gemini. My expertise in applying graph mining for data curation has been recognized with multiple internal "Tech and Research Impact Awards."
Prior to this, I was a tech lead for Kartta Labs (github.com/kartta-labs), an open-source Google Research project aimed at reconstructing a 4D map of the world. Launched as re.city in 2020, the project concluded in 2023 (archival videos available on YouTube).
My PhD research at the University of Southern California, advised by Patrick Lynett and Cyrus Shahabi, resulted in Celeris. This open-source software has democratized high-performance computing in coastal engineering, used by thousands of researchers and engineers across over 50 countries, and has seen its user manual translated into Spanish, Farsi, and Italian by independent users.
Seeking AI guidance? I'm available to discuss data curation and RAG for LLMs and AI more broadly.
EMPLOYMENT
Google AI
Google AI
Senior Software Engineer, 2018 - Present
2021-Now: I am leading the Contextual Sampling team of Google, developing and applying scalable graph-based techniques on applications ranging from Gemini to Search AI Overview and YouTube.
2018-2020: I was a tech lead in github.com/kartta-labs, an open-source project for unrendering historical maps and reconstructing 4D map of the world. We used artificial intelligence and crowdsourcing to tackle tasks such as 3D reconstruction, georectification, vectorization, etc. Have a look at it here: http://re.city.
Celeria Labs, LLC
Celeria Labs, LLC
Founder, 2020 - Present
Celeria Labs specializes in computational hydrodynamics. Its major product is Celeris, a simulation software for nearshore waves. Learn more at celerialabs.com.
Niantic, Inc.
Niantic, Inc.
Software Engineering Intern, Summer 2018
I was the very first Intern of Niantic, the company behind Pokémon Go. I contributed to the launch of Ingress Prime as part of the core Unity3D engineering team.
Google
Google
Software Engineering Intern, Winter 2018
As a member of the core engineering team of Code Jam, I significantly contributed to the launch of Code Jam 2018 competition on the new platform and received a shiny spot bonus for the “heroic launch”.
Google
Google
Software Engineering Intern, Summer 2017
As a member of the Brand Lift team, I scaled up the ad quality measurement tools both in time and space, from spontaneous runs in specific regions to continuous monitoring in several countries.
EDUCATION
University of Southern California
University of Southern California
Master of Science in Computer Science 2016
Thesis: Efficient Geospatial Crowdsourcing for Post-Disaster Decision Making [CONF. PAPER]
Amirkabir University of Technology (Tehran Polytechnic)
Amirkabir University of Technology (Tehran Polytechnic)
Master of Science in Computational Hydraulics 2013
Thesis: Curvilinear Smoothed Particle Hydrodynamics [JOURNAL PAPER]
Amirkabir University of Technology (Tehran Polytechnic)
Amirkabir University of Technology (Tehran Polytechnic)
Bachelors of Science in Civil Engineering 2010
Thesis: Shear Force Distribution in Open Channels [JOURNAL PAPER]
Gemini Team, et al. "Gemini: a family of highly capable multimodal models." arXiv preprint arXiv:2312.11805 (2023).
Tavakkol, S., Son, S., & Lynett, P. (2021). Adaptive third order Adams-Bashforth time integration for extended Boussinesq equations. Computer Physics Communications, 265, 108006.
Tavakkol, S., Shahabi, C., Han, F., & Kiveris, R. (2020, December). Piaget: A Probabilistic Inference Approach for Geolocating Historical Buildings. In 2020 IEEE International Conference on Big Data (Big Data) (pp. 971-978). IEEE.
Tavakkol, S., & Lynett, P. (2020). Celeris Base: An interactive and immersive Boussinesq-type nearshore wave simulation software. Computer Physics Communications, 248, 106966.
Tavakkol, S., Chiang, Y. Y., Waters, T., Han, F., Prasad, K., & Kiveris, R. (2019, November). Kartta labs: Unrendering historical maps. In Proceedings of the 3rd ACM SIGSPATIAL International Workshop on AI for Geographic Knowledge Discovery (pp. 48-51).
Alialy, R., Tavakkol, S., Tavakkol, E., Ghorbani-Aghbologhi, A., Ghaffarieh, A., Kim, S. H., & Shahabi, C. (2018). A Review on the Applications of Crowdsourcing in Human Pathology. Journal of Pathology Informatics, 9. [PDF]
Tavakkol, S., & Lynett, P. (2017). Celeris: A GPU-accelerated open source software with a Boussinesq-type wave solver for real-time interactive simulation and visualization. Computer Physics Communications, 217, 117-127. {PDF] [WEBSITE]
Tavakkol, S., Zarrati, A. R., & Khanpour, M. (2017). Curvilinear smoothed particle hydrodynamics. International Journal for Numerical Methods in Fluids, 83(2), 115-131. [PDF]
Tavakkol, S., To, H., Kim, S. H., Lynett, P., & Shahabi, C. (2016). An entropy-based framework for efficient post-disaster assessment based on crowdsourced data. In Proceedings of the 2nd ACM SIGSPATIAL International Workshop on the Use of GIS in Emergency Management (EM-GIS '16). ACM, San Francisco, CA, USA. [PDF]
Tavakkol, S., Alapour, F., Kazemian, A., Hasaninejad, A., Ghanbari, A., & Ramezanianpour, A. A. (2013). Prediction of lightweight concrete strength by categorized regression, MLR and ANN. Computers and Concrete, 12, 151-167. [PDF]
Google Sites
Report abuse