About Me
I am currently a research scientist and manager at Snap Research, leading a team of scientists, engineers and interns and collaborators on fundamental and applied research initiatives in user modeling and personalization across Snapchat. I am broadly interested in advancing the state-of-the-art in machine learning algorithms and applications the realm of large-scale structured data -- such as graph and sequential representations -- and generative recommendation systems. I am especially in applications to recommendation systems and trust and safety problems at scale. My work broadly spans the machine learning, data mining, and software engineering domains.
Before I joined Snap, I completed my PhD in the Computer Science
Department at Carnegie Mellon University, where I
worked on discovering and modeling various types of abusive online behaviors in large networks. I was very fortunate to have been advised by Christos Faloutsos.
Prior to this, I received my B.S. in Computer Science from the Department
of Computer Science at North Carolina State University. There, I worked with Nagiza Samatova on
reduction, indexing and storage systems for large-scale scientific data.
If you would like to learn more about my work, please take a look at my publications
and CV. Please reach out to me via my personal e-mail for consulting inquiries.
We're hiring multiple
full-time and
intern Research Scientist roles for 2026 working on language models, information retrieval and user personalization. Please reach out and apply if you have prior work in these areas.
Recent News
- November 2025: Sharing a new preprint on model-scaling behavior in generative recommendation methods, which shows scaling limitations in existing semantic ID-based methods.
- October 2025: Excited to share two new works at CIKM 2025 on generative recommendation, covering the newest open-source reproducibility tooling (GRID) and meta-item embeddings for cold-start learning.
- October 2025: Excited to share our new work at LoG 2025 on GNN distillation to MLPs, which shows that stronger models aren't always stronger teachers.
- May 2025: We have several works at KDD 2025 on graph neural networks (GiGL, our library to scale GNNs at Snap, and a corresponding tutorial), and recommendation systems (improved self-attention for cross-domain recommendation, and optimization in collaborative filtering)!
- May 2025 Excited to share we have two works accepted at ICML 2025 on rethinking historical embeddings in GNN scaling and hyperbolic geometry for backwards-compatible learning.
- Apr 2025: Excited to share our work on learning cross-domain user representations at Snapchat was accepted to SIGIR 2025 -- see you in Padua!
- Mar 2025: Excited to share our workshop proposal on GenAI for Recommender Systems and Personalization was accepted to KDD 2025 -- see you in Toronto!
- Mar 2025: Sharing an exciting new preprint talking about our efforts to productionize and apply large-scale GNNs at Snap.
- Feb 2025: Our work introducing the first multimodal graph learning benchmark was accepted to CVPR 2025 -- see you in Nashville!
- Jan 2025: Our works on graph-clustering for parameter compression and interpreting collaborative filtering through matrix rank were accepted to WWW 2025 -- see you in Sydney!
- Sep 2024: Our work on test-time message passing for collaborative filtering was accepted to NeurIPS 2024 -- see you in Vancouver!
- Aug 2024: Check our some interesting preprints on neural scaling laws with graph models and the first multimodal graph benchmark!
- July 2024: Check out a preprint of NodeDup, a super simple approach for drastically improving graph link prediction performance with data augmentation!
- May 2024: Our work on language models for content moderation was accepted to ACL 2024 -- see you in Bangkok!
- April 2024: We have multiple papers at ICML 2024 on exploring graph foundation models and adapting language models for graph tasks -- see you in Vienna!
- March 2024: We'll present an iteration of our popular large-scale graph neural networks tutorial at SDM 2024 -- see you in Houston!
- February 2024: Excited to give a tutorial at AAAI 2024 on large-scale graph neural networks -- see you in Vancouver!
- February 2024: We will present multiple works at ICLR 2024 on graph neural networks, specifically around demystifying link prediction performance and conceptualizing graph learning without explicit graph models -- see you in Vienna!
- October 2023: We have multiple papers at NeurIPS 2023, on cold-start graph learning, evaluation pitfalls in link prediction and graph structural disparity -- see you in New Orleans!
- August 2023: Excited to be deeply involved at KDD 2023, participating in panels, workshops, tutorials and paper presentations; come say hello in Long Beach!
- August 2023: Honored that our work Graph Data Augmentation for Graph Machine Learning: A Survey was featured in the IEEE Data Engineering Bulletin edition on Graph Neural Networks.
- July 2023: It's a busy week for our team at Snap! Several of us will be at SIGIR 2023, presenting our work on Embedding-based Retrieval in Friend Recommendation, and several others will be at ICML 2023, presenting our work on Linkless Link Prediction via Relational Distillation.
- June 2023: We have several papers at KDD 2023, on graph anomaly detection, clustering-based graph self-supervised-learning, and scaling up graph neural networks -- see you in Long Beach!
- May 2023: Our paper Are Message Passing Neural Networks Really Helpful for Knowledge Graph Completion? was accepted to ACL 2023 -- see you in Toronto!
- May 2023: Our paper Linkless Link Prediction via Relational Distillation was accepted to ICML 2023 -- see you in Honolulu!
- Apr 2023: Our tutorial Large-Scale Graph Neural Networks: The Past and New Frontiers was accepted to KDD 2023 -- see you in Long Beach!
- Mar 2023: Our paper Embedding-based Retrieval in Friend Recommendation was accepted to SIRIP 2023 -- see you in Taipei!
- Feb 2023: Our tutorial Augmentation Methods for Graph Learning was accepted to SDM 2023 -- see you in Minneapolis!
- Jan 2023: Our paper MLPInit: Embarrassingly Simple GNN Training Acceleration with MLP Initialization was accepted to ICLR 2023 -- see you in Kigali!
- Jan 2023: Our paper Multi-task Self-supervised Graph Neural Networks Enable Stronger Task Generalization was accepted to ICLR 2023 -- see you in Kigali!
- Jan 2023: Our paper Link Prediction with Non-Contrastive Learning was accepted to ICLR 2023 -- see you in Kigali!
- Jan 2023: Our paper Empowering Graph Representation Learning with Test-Time Graph Transformation was accepted to ICLR 2023 -- see you in Kigali!
- Nov 2022: Our paper Flashlight: Scalable Link Prediction with Effective Decoders was accepted to LoG 2023 -- see you over Zoom!
- Nov 2022: Our paper Graph Explicit Neural Networks: Explicitly Encoding Graphs for Efficient and Accurate Inference was accepted to WSDM 2023 -- see you in Singapore!
- Oct 2022: Excited to serve as a panelist at the Stanford Graph Learning Workshop 2022 -- see you in Stanford, CA!
- Sep 2022: Our paper Explaining Graph Neural Networks with Structure-Aware Cooperative Games was accepted to NeurIPS 2022 -- see you in New Orleans!
- Sep 2022: Our paper A Practical, Progressively Expressive Graph Neural Network was accepted to NeurIPS 2022 -- see you in New Orleans!
- Aug 2022: Our paper Imbalanced Graph Classification via Graph-of-Graph Neural Networks was accepted to CIKM 2022 -- see you in Atlanta!
- Aug 2022: Honored to give a keynote and serve on a panel at the KDD 2022 Deep Learning on Graphs (DLG) Workshop!
- Aug 2022: Excited to co-organize the KDD 2022 Mining and Learning with Graphs (MLG) Workshop!
- July 2022: Our paper Predicting Future Location Categories of Users in a Large Social Platform was accepted to ICWSM 2023 -- see you in Cyprus!
- May 2022: Excited to moderate a panel at the TigerGraph AI Summit on graph deep learning libraries!
- May 2022: Honored to serve as an invited speaker and panelist at the Knowledge Graph Conference 2022, where I'll talk about advances in industrial-scale graph ML!
- Apr 2022: Looking forward to giving an invited talk at the UC Riverside CSE Colloqium on our work about Scaling Up Graph Neural Networks at Snap!
- Mar 2022: Honored to receive the Outstanding Service Award at WSDM 2022!
- Jan 2022: Our paper Sunshine with a Chance of Smiles: How does Weather Impact Sentiment on Social Media? was accepted to ICWSM 2022 -- see you over Zoom!
- Feb 2022: Excited to organize WSDM Cup 2022 on Feb 25th, and see theory meet practice across tasks like user retention prediction, temporal link prediction and cross-market recommendation!
- Feb 2022: Excited to give a keynote at the WSDM Machine Learning on Graphs (MLoG) Workshop and talk about some of our recent advances in scaling up graph neural network training and inference at Snap!
- Jan 2022: Our paper Graph-less Neural Networks: Teaching Old MLPs new Tricks via Distillation was accepted to ICLR 2022 -- see you over Zoom!
- Jan 2022: Our paper Graph Condensation for Graph Neural Networks was accepted to ICLR 2022 -- see you over Zoom!
- Jan 2022: Our paper From Stars to Subgraphs: Uplifting any GNN with Local Structure Awareness was accepted to ICLR 2022 -- see you over Zoom!
- Jan 2022: Our paper Is Homophily a Necessity for Graph Neural Networks? was accepted to ICLR 2022 -- see you over Zoom!
- Jan 2022: Our paper Automated Self-Supervised Learning for Graphs was accepted to ICLR 2022 -- see you over Zoom!
- Oct 2021: Excited to talk about some of our recent advances in scarce-label graph learning at the IIIT Delhi Laboratory for Computational Social Systems Seminar Series as an invited speaker.
- Oct 2021: Our paper Ranking Friend Stories on Social Platforms with Edge-Contextual Local Graph Convolutions was accepted to WSDM 2022 -- see you in Phoenix!
- Oct 2021: Our paper Attributed Graph Modeling with Vertex Replacement Grammars was accepted to WSDM 2022 -- see you in Phoenix!
- Oct 2021: Our paper Finding a Concise, Precise and Exhaustive Set of Near Bi-Cliques in Dynamic Graphs was accepted to WSDM 2022 -- see you in Phoenix!
- Aug 2021: Our paper Action Sequence Augmentation for Early Graph-based Anomaly Detection was accepted to CIKM 2021 -- see you over Zoom!
- Aug 2021: Our paper A Unified View on Graph Neural Networks as Graph Signal Denoising was accepted to CIKM 2021 -- see you over Zoom!
- Aug 2021: Our paper Niche Detection in User Content Consumption Data was accepted to CIKM 2021 -- see you over Zoom!
- Aug 2021: Our paper A Synergistic Approach for Graph Anomaly Detection with Pattern Mining and Feature Learning was accepted to TNNLS 2021!
- Jun 2021: I will be co-organizing WSDM Cup at WSDM 2022 -- the call for task proposals is out; looking forward to your amazing submissions!
- Apr 2021: Our paper FairOD: Fairness-aware Outlier Detection was accepted to AIES 2021 -- see you over Zoom!
- Mar 2021: I will be co-organizing the Misinformation and Misbehavior (MIS2) workshop at KDD 2021 this year -- see you over Zoom!
- Mar 2021: I will be a keynote speaker at the ML in Finance (MLF) workshop at KDD 2021 this year -- see you over Zoom!
- Mar 2021: Our paper Online Communication Shifts in the Midst of the Covid-19 Pandemic: A Case Study on Snapchat
was accepted to ICWSM 2021 -- see you (hopefully) in Atlanta, GA!
- Mar 2021: Our paper CEAM: The Effectiveness of Cyclic and Ephemeral Attention Models of User Behavior on Social Platforms
was accepted to ICWSM 2021 -- see you (hopefully) in Atlanta, GA!
- Dec 2020: Our paper AdverTiming Matters: Examining User Ad Consumption for Effective Ad Allocations on Social Media
was accepted to CHI 2021 -- see you over Zoom!
- Dec 2020: Our exciting new work FairOD: Fairness-aware Outlier Detection is now available as a preprint!
- Dec 2020: Our paper Data Augmentation for Graph Neural Networks was accepted to AAAI 2021 -- see you over Zoom!
- Oct 2020: Our exciting new work A Unified View on Graph Neural Networks as Graph Signal Denoising is now available as as preprint!
- Sep 2020: Our paper The Devil is in the Details: Evaluating Limitations of Transformer-based Methods for Granular Tasks was accepted to COLING 2020 -- see you over Zoom!
- Jul 2020: Our paper Social Factors in Closed-Network Content Consumption was accepted to CIKM 2020 -- see you over Zoom!
- Jul 2020: Our paper Identifying Misinformation from Website Screenshots was accepted to ICWSM 2021 -- see you (hopefully) in Venice!
- Jul 2020: Our paper Scale-Free, Attributed and Class-Assortative Graph Generation to Facilitate Introspection of Graph Neural Networks was accepted to KDD MLG 2020 -- see you over Zoom!
- Jun 2020: Our exciting new work Data Augmentation for Graph Neural Networks is now available as a preprint!
- Jun 2020: Our paper Semi-Supervised Multi-aspect Misinformation Detection with Hierarchical Joint Decomposition was accepted to ECML-PKDD 2020 -- see you over Zoom!
- May 2020: Our paper Knowing your FATE: Friendship, Action and Temporal Explanations for User Engagement Prediction on Social Apps was accepted to KDD 2020 -- see you over Zoom!
- Feb 2020: I will be chairing the CyberSafety
workshop at TheWebConf 2020 --
see you over Zoom!
- Oct 2019: I will be a keynote speaker and panelist at the Doctoral Consortium at ICDM 2019 -- see you in Beijing!
- Jul 2019: I will be giving a keynote talk on Outlier Detection for Mining Social Misbehavior at the Learning and Mining for Cybersecurity (LEMINCS) workshop at KDD 2019 -- see you in Anchorage!
- Jul 2019: Our paper SliceNDice: Mining Suspicious Multi-attribute Entity Groups with Multi-view Graphs was accepted to DSAA 2019 -- see you in Washington, DC!
- Jul 2019: Our paper FARE: Schema-Agnostic Anomaly Detection in Social Event Logs was accepted to DSAA 2019 -- see you in Washington, DC!
- Jul 2019: I will be chairing WSDM Cup at WSDM 2020 -- see you in Houston!
- Jun 2019: Our paper Characterizing and Detecting
Livestreaming
Chatbots
was accepted to ASONAM 2019 -- see you in Vancouver!
- Apr 2019: Our paper Modeling
Dwell Time Engagement on Visual Multimedia was accepted to KDD 2019 -- see you in
Anchorage!
- Mar 2019: Our CHI 2019 paper on Impact
of Contextual Factors on Public Snapchat Sharing
won a Best Paper Honorable
Mention award!
- Dec 2018: Our paper Impact of
Contextual Factors on Public Snapchat Sharing
was accepted to CHI 2019 -- see
you in Glasgow!
- Dec 2018: Our work on false information with Vagelis Papalexakis was featured on
the NVIDIA AI
Podcast!
- Nov 2018: I will be chairing the CyberSafety
workshop at WWW 2019 --
see you in San Francisco!
- Oct 2018: I will be chairing the ASONAM 2019 Industrial Track --
see you in
Vancouver!
- Sep 2018: I will be giving a keynote talk on Lessons I learned during my PhD at the PhD Forum at ECML-PKDD 2019 -- see
you in Dublin!
- Aug 2018: Our work on false information with Vagelis Papalexakis is featured in a
Digital
Trends article.
- Jul 2018: I will be giving a keynote talk on Outlier Detection for Mining Social Misbehavior at the Outlier Detection
De-constructed (ODD)
workshop at KDD 2018 -- see you in London!
- Jun 2018: Our paper Did We
Get It Right? Predicting Query Performance in E-commerce Search was accepted to the
eCom Workshop at SIGIR 2018!
- Jun 2018: Our paper Beyond Outlier
Detection: LookOut for Pictorial Explanation was accepted to ECML-PKDD 2018 -- see
you in Dublin!
- Jun 2018: Our paper Semi-supervised
Content-based Detection of Misinformation via Tensor Embeddings was accepted to
ASONAM 2018 -- see you in Barcelona!
- Apr 2018: Srijan
Kumar and I
published
a comprehensive survey on characterizing and detecting false information, False Information on Web and Social Media: A
Survey. Check it out!
- Mar 2018: I will be giving a keynote talk on Anomaly Detection on Large Social Graphs at the CyberSafety workshop at WWW 2018 --
see you in Lyon!
- Jan 2018: Our paper Reducing Large Graphs
to Small Supergraphs: A Unified Approach will be published in SNAM 2018!
- Dec 2017: I have joined Snap Inc as a Research Scientist in Los
Angeles, CA!
- Oct 2017: I successfully defended my thesis, Anomaly Detection in Large Social Graphs, from Carnegie
Mellon University's School of Computer Science, advised by Christos Faloutsos.