curriculum vitae

General Information

Full Name Otilia Stretcu
Contact otiliastr [at] gmail [dot] com
Languages Romanian (native), English (fluent), Spanish (beginner), German (beginner), Swedish (beginner)

Research Areas

  • I work on machine learning methodology—primarily around vision-language models, active learning, and safety—with applications in various areas including computer vision and natural language processing. My current research focuses on two key problems:
  • 1) Enabling domain experts to effectively build models without requiring AI expertise. I work on automating the process of mining the right kinds of data and iteratively improving the model, which spans multiple areas including modeling, active learning and distillation.
  • 2) Improving AI Trust & Safety models to make the internet safer, by leveraging LLM and VLM capabilities to detect and counteract malicious activities. This involves fundamental research aimed at expanding the reasoning capabilities of LLMs over multimodal data, to better identify sophisticated harmful content and malicious behaviors.
  • During my PhD I also worked on curriculum learning, semi-supervised learning, multi-task learning, and graph-based problems. I am also passionate about using machine learning methods for health and sciences.

Education

  • 2015 - 2021
    Carnegie Mellon University – Ph.D. in Machine Learning
    • Co-advised by Prof. Tom Mitchell and Dr. Barnabàs Pòczos
    • GPA: 4.0 (4.0 scale)
    • Thesis: Curriculum Learning
    • Thesis committee: Tom Mitchell, Barnabàs Pòczos, Ruslan Salakhutdinov, Rich Caruana
  • 2015 - 2017
    Carnegie Mellon University – M.S. in Machine Learning
    • Co-advised by Prof. Tom M. Mitchell and Dr. Barnabàs Pòczos
    • GPA: 4.0 (4.0 scale)
    • Thesis: Understanding the Neural Basis of Speech Production Using Machine Learning
    • Master’s degree requirements completed while working towards obtaining my Ph.D.
  • 2014 - 2015
    University of Cambridge – Master of Philosophy (M.Phil.) in Advanced Computer Science
    • Advised by Prof. Pietro Lió
    • Thesis: Machine Learning Methods for Computational Microscopy
    • Grade: Pass with Distinction
  • 2012 - 2013
    Linköping University – Erasmus Exchange Student
    • I spent the third year of my undergraduate studies as an Erasmus exchange student at Linköping University, Sweden
  • 2010 - 2014
    Politehnica University of Timisoara – B.Eng. in Computer Science and Information Technology
    • GPA: 9.98 (scale 10.0)
    • 1st out of 140 students

Work Experience

  • 2023 - now
    Senior Research Scientist at Google AI
    • Mountain View, CA, USA
  • 2021 - 2023
    Research Scientist at Google AI
    • Mountain View, CA, USA
  • Summer 2018
    &
    Spring 2019
    Software Engineering Intern at Google Research
    • Expander team in Google Research , Mountain View, CA, USA
    • Research on deep learning models for graph-based semi-supervised learning, published at NeurIPS 2019.
  • Summer 2016
    Software Engineering Intern at Google X
  • Summer 2014
    Software Developer Intern at Microsoft

Other Research Experience

  • 2014 - 2015
    Independent Research Project in Computer Vision
    • Advised by Dr. Marius Leordeanu
    • Unsupervised object discovery in video based on multiple frames matching. We also proposed a fast method for detecting the main object of interest in a video, titled VideoPCA. Published at BMVC 2015.
  • Summer 2013
    Research Internship in Machine Learning at EPFL
    • Laboratory for Probabilistic Machine Learning at École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland
    • Advised by Dr. Matthias Seeger
    • Used topic models to explore the correlation between social media messages from Twitter and the location of the users, with applications to user profiling, topic tracking and content recommendation. I was responsible with applying various machine learning models and parallelizing the code in order to scale well.
  • Summer 2013
    Research for Undergraduates Program
    • Politehnica University of Timisoara, Romania
    • Advised by Prof. Emilia Petrisor
    • Implemented algorithms for spectral clustering of nodes in a graph, based on minimum graph cut, with applications to data mining and statistics, such as clustering information from documents on the web and medical images segmentation.
Note: for further details on my research, please see the publications page.

Honors and Awards

  • Fellowships
  • Scholarships
  • Awards
    • Google Research Tech Impact Award for developing cutting-edge AI for improving trust & safety across Google products (2024)
    • Google Research Tech Impact Award for developing machine learning technology that enables users without AI expertise to effortlessly train AI models (2023)
    • Google Ads Tech Impact Award for developing ads safety technology (2024, 2023, 2022)
    • Best poster award at the Eastern European Machine Learning Summer School in Bucharest, Romania (2019)
    • Machine Learning Department Teaching Assistant Award (2018)
    • Carnegie Mellon University Neurohackathon: 2nd place (2017)
    • KTH University Programming Challenge, Sweden: Top 10 contestants (2013)
    • ACM International Collegiate Programming Contest (ACM-ICPC): Honorable Mention in Southeastern European Regional (2013, 2012, 2011)
    • Microsoft Imagine Cup:
      • Top 13 in the World Finals (2012)
      • 1st team in the Romanian National Finals (2012)
    • Romanian National Olympiad in Informatics:
      • Gold Medal: 2008
      • Bronze Medal: 2010
      • 1st Place: 2004
      • 2nd Place: 2005
      • Honorable Mention: 2010, 2008, 2007, 2003
    • Kangaroo International Mathematical Competition: 2nd prize in Romanian National Finals (2009, 2010)

Invited talks

  • ○ Invited talk at the ICLR workshop I Can’t Believe It’s Not Better: Challenges in Applied Deep Learning (2025).
  • ○ Lecture on "Jointly modeling images and text" at the Polytechnic University of Bucharest, Romania, as a guest lecturer in the Computer Vision class, part of the Master’s program in AI (2022).
  • ○ Invited talk at the Quantitative Research Colloquium (QRC) hosted by Morgan Stanley (2021).
  • ○ Invited talk at Health@Scale on Graph Agreement Models for Semi-Supervised Learning (2020).
  • ○ Represented CMU at the MIDAS Data Science Annual Symposium at the University of Michigan (2019).
  • ○ Talk at the CMU AI Seminar on Contextual Parameter Generation for Knowledge Graph Link Prediction (2019).

Teaching Experience

  • Spring 2018
    Teaching Assistant for Graduate Machine Learning
  • Fall 2017
    Teaching Assistant for Topics in Deep Learning
  • 2013 - 2014
    Teaching algorithms for competitive programming
    • Co-organized a competitive programming seminar at Politehnica University of Timisoara for university and high-school students interested to train for algorithmic competitions (e.g. ACM-ICPC, informatics olympiad).
    • Taught algorithms and data structures used in competitive programming, designed and solved practice problems and internal competitions.

Technical Projects

  • LiveX Learning Platform
    • Tutoring system for kindergarten and school children based on a software platform that runs in the cloud, Windows Phone 7 devices and a set of electronic learning cubes called “IQubes” (our hardware invention).
    • Project proposed by our team, called IQube, that competed in the world finals of the Microsoft Imagine Cup competition.
  • Face and Hand Gesture Recognition for Human - Computer Interaction
    • Framework for C++ developers to extend their graphical user interfaces with more natural means of communication.
    • Works in real-time using a computer web camera.
  • Public Transport Route Recommendation
    • Python application for the Timisoara city public transport system using real-time information from GPS devices installed on public transport vehicles.
    • Overlays optimal routes suggestions on Google Maps (before they supported such a feature).
  • Handwritten digits recognition
    • C library implementing various linear algebra methods.

Leadership and Volunteering Activities

  • 2019 - now
    • Mentorship:
    • - Mentor in the Mind the gap program organized by Google, which aims to increase representation of girls in tech (2022)
    • - Mentor for the Carnegie Mellon University AI mentoring program (2019 - 2021)
    • - Mentor for first year PhD students at Carnegie Mellon University (2019 - 2021)
  • 2018 - now
    • Program Committees: NeurIPS (2023, 2022, 2021, 2020), ICML (2019), AISTATS (2020, 2019), ICLR (2020, 2018), ICLR-LLD (2019), PLOS ONE (2019), ICML-GRL (2020), AAAI (2021), Google Research Scholar Program (2023, 2024).
  • 2019
  • 2018 - 2021
    • Founding member of the AI+ Club at Carnegie Mellon University (CMU).
  • 2018 - 2019
    • Treasurer of the Romanian Students Association at CMU.
  • 2016 - 2021
    • Member of the Doctoral Review Committee of the Machine Learning Department at CMU.
  • 2016 - 2021
    • Member of the Education Review Committee of the Machine Learning Department at CMU, which aims to improve the PhD program.
  • 2016 - 2018
    • President of the Romanian Students Association at CMU.
  • 2011 - 2012
    • Student representative in the faculty leadership board at Politehnica University of Timisoara.
  • 2010 - 2011
    • Volunteer for AIESEC, international youth organization.
  • 2010 - 2012
    • Volunteer for Liga AC, student organization at Politehnica University.

Other Interests

  • Sports: squash, volleyball, tennis, climbing, hiking.
  • Hobbies: traveling, reading, arts and crafts, learning languages on Duolingo, GeoGuessr