Vinitra Swamy

Vinitra Swamy

Scholé AI · Harvard · vinitra@berkeley.edu

Hi there! I'm an AI researcher, CEO and co-founder of the AI for education spin-off Scholé AI. I also teach Harvard's Agentic AI Intensives alongside some stellar faculty.

I earned my PhD in Computer Science at in 2025, co-advised by Prof. Tanja Käser and Prof. Martin Jaggi. My PhD research was awarded the Patrick Denantes Prize for EPFL's best CS thesis, and I was also fortunate to be named a Rising Star in Data Science by Stanford University.

Before Switzerland, I spent two years at Microsoft AI as a lead engineer for the Open Neural Network eXchange (ONNX) project, collaborating with more than 30 leading tech companies, including Nvidia, Intel, Amazon, and Meta, to help establish an industry standard for AI interoperability.

My claim to fame (😂) is graduating at 20 as the youngest M.S. in Computer Science recipient in UC Berkeley's history. Since then, I've taught AI and machine learning at UC Berkeley and University of Washington, and spent summers at , , , and .

I love people, data, and working on exciting problems at the intersection of the two:

  • ML for education (personalized learning, autograding, knowledge tracing)
  • Explainable and interpretable AI
  • Generalized learning (transfer learning, multimodal learning)

Thank you for taking time from your day to find out what I do with mine!

In the press
Scholé AI co-founders Vinitra Swamy and Paola Mejia
ACE Ventures · 2026

Scholé AI raises $3M to transform workforce learning in the age of AI
Vinitra celebrating her thesis defense at EPFL
EPFL News · 2025

Patrick Denantes Memorial Prize for EPFL's best computer science thesis
Ambitious x Driven feature on Vinitra
Ambitious x Driven · 2024

Google, LinkedIn, Microsoft, and now EPFL PhD: Vinitra's incredible journey
Vinitra in Lausanne, Switzerland
G-Research · 2024

G-Research PhD Prize winner at EPFL

Selected Research

✨🎓 My PhD thesis on "A Human-Centric Approach to Explainable AI for Personalized Education" is now available!

For a full list of publications, please visit my Google Scholar page.

During my PhD, I was fortunate to be recognized by Stanford, UCSD and UChicago as a "Rising Star in Data Science", and awarded the GResearch PhD Award, the IC Distinguished Service Award three times, and most recently the Patrick Denantes Memorial Prize for EPFL's best CS thesis.

500+
2026ACM L@S

Turning 500+ Students into Teachers

Chenyang Wang, Christopher Petrie, Miltiadis Stouras, Nicolas Ettlin, Amaury George, Paola Mejia-Domenzain, Vinitra Swamy, Tanja Käser, Ola Svensson
A semester-long study of an AI teachable agent in an undergraduate algorithms course. We present Explique, a platform integrating the teachable agent Algorithm Apprentice to operationalise learning-by-teaching at scale, and report an 11-week field deployment with 546 students and 3,809 student-agent dialogues. Using generalised linear mixed-effects models, we find that explanation-oriented dialogue behaviours (elaboration, showing reasoning) are associated with fewer incorrect quiz submissions, whereas reuse of externally sourced content is associated with slightly more repeated attempts.
Paper
AI ⇄ 🧑
2026JCAL

Who Gives Feedback Matters

Tanya Nazaretsky, Paola Mejia-Domenzain, Vinitra Swamy, Jibril Frej, Tanja Käser
A journal extension of our ECTEL best paper, examining student biases towards human and AI-generated formative feedback. Students evaluate feedback in authentic educational settings both before and after the provider's identity is disclosed, isolating the effect of attribution on perceived quality. Journal of Computer Assisted Learning, 42(1), e70153.
Paper News
🌍 XAI
2026AIED

User-Driven XAI in Education & Human Agency

Hasan Abu-Rasheed, Jakub Kuzilek, Mutlu Cukurova, Hassan Khosravi, Luc Paquette, Tanja Käser, Benjamin Paaßen, Christian Weber, Vinitra Swamy, Qianhui Sophie Liu
A community position piece on shifting explainable AI in education from system-driven to user-driven explanation, and what that shift implies for human agency — enabling teachers, learners, and institutions to understand, contest, and actively shape the AI systems affecting their educational experiences.
Paper
UNESCO
2026AIED

Scenario-Based Learning Through and With AI

Ella Hamonic, Candy Lugaz, Annina Demirag, Rémi Sharrock, Agustina Thailinger, Maria Victoria Picchio, Vinitra Swamy, Paola Mejia-Domenzain
A collaboration with IIEP-UNESCO and Télécom Paris on evidence-informed simulations for education leaders. AI generates context-aware scenarios, delivers the simulation, prompts reflection, and provides coaching-style feedback grounded in the school leadership and educational planning literature — placing leaders in realistic school situations where they analyse challenges, take decisions, and justify their reasoning.
Paper
SCRIBE structured chain reasoning with tool calls
2025EMNLP

SCRIBE: Structured Chain Reasoning with Tool Calling

Fares Fawzi, Vinitra Swamy, Dominik Glandorf, Tanya Nazaretsky, Tanja Käser
Real-world deployment of student feedback models faces three key challenges: privacy, limited compute, and pedagogical validity. We introduce SCRIBE, a framework for multi-hop, tool-augmented reasoning that combines domain-specific tools with a self-reflective inference pipeline supporting iterative reasoning, tool use, and error recovery. We distil these capabilities into 3B and 8B models via two-stage LoRA fine-tuning on synthetic GPT-4o data. Evaluation with a human-aligned GPT-Judge and a user study with 108 students shows 8B-SCRIBE achieves comparable or superior quality to much larger models on relevance and actionability, while being perceived on par with GPT-4o and Llama-3.3 70B.
Paper Pre-Print Code
InterpretCC feature gating and group routing architectures
2025ICLRTop 5%

InterpretCC: User-Centric Interpretability via Mixture-of-Experts

Vinitra Swamy, Syrielle Montariol, Julian Blackwell, Jibril Frej, Martin Jaggi, Tanja Käser
We present InterpretCC (interpretable conditional computation), a family of intrinsically interpretable neural networks at a unique point in the design space that optimizes for ease of human understanding and explanation faithfulness, while maintaining comparable performance to state-of-the-art models. InterpretCC achieves this through adaptive sparse activation of features before prediction, allowing the model to use a different, minimal set of features for each instance. We extend this into an interpretable, global mixture-of-experts model that lets users specify topics of interest, discretely separates the feature space into topical subnetworks, and adaptively activates them for prediction. Across text, time series and tabular data, InterpretCC matches non-interpretable baselines and outperforms intrinsically interpretable ones. In a user study with 56 teachers, its explanations rate higher on actionability and usefulness.
Paper OpenReview Pre-Print Code
iLLuMinaTE four-stage LLM-XAI pipeline
2025AAAI

iLLuMinaTE: From Explanations to Action

Vinitra Swamy*, Davide Romano*, Bhargav Srinivasa Desikan, Oana-Maria Camburu, Tanja Käser
We introduce iLLuMinaTE, a zero-shot, chain-of-prompts LLM-XAI pipeline inspired by Miller's cognitive model of explanation, designed to deliver theory-driven, actionable feedback to students in online courses. It navigates three stages — causal connection, explanation selection, and explanation presentation — with variations drawing from eight social science theories. We evaluate 21,915 natural language explanations extracted from three LLMs (GPT-4o, Gemma2-9B, Llama3-70B) with three XAI methods (LIME, Counterfactuals, MC-LIME) across three diverse online courses, including a real-world preference study with 114 university students and a novel actionability simulation. Students prefer iLLuMinaTE explanations over traditional explainers 89.52% of the time.
Paper Pre-Print Code
JAIR
2025JAIR

Viewpoint: Future of Human-Centric XAI

Vinitra Swamy, Jibril Frej, Tanja Käser
Current approaches in human-centric XAI (predictive tasks in healthcare, education, or personalized ads) tend to rely on a single explainer — concerning given systematic disagreement between explainability methods applied to the same points and underlying black-box models. We propose shifting from post-hoc explainability to designing interpretable neural network architectures. We identify five needs of human-centric XAI (real-time, accurate, actionable, human-interpretable, and consistent) and propose two schemes for interpretable-by-design neural network workflows. We postulate that the future of human-centric XAI is neither in explaining black-boxes nor in reverting to traditional interpretable models, but in neural networks that are intrinsically interpretable.
Paper Pre-Print
🤝 Trust
2025Computers & Education: AI

Measuring Student Trust in AI-Powered EdTech

Tanya Nazaretsky, Paola Mejia-Domenzain, Vinitra Swamy, Jibril Frej, Tanja Käser
Trust is a key factor determining whether students engage with AI-powered educational tools. This paper presents the development and validation of an instrument for measuring student trust in AI-powered educational technology, grounded in a multidimensional conceptualisation of trust. Validated with a large sample of university students, it offers the learning analytics and AIED communities a practical tool for assessing and improving trust in AI-based learning systems.
Paper
🎓 PhD
2025EPFL Thesis🏆 Best Thesis

A Human-Centric Approach to Explainable AI for Personalized Education

Vinitra Swamy — advised by Tanja Käser and Martin Jaggi
This thesis brings human needs to the forefront of XAI research, grounded in the concrete use case of personalized learning and teaching. We frame the contributions along two verticals: technical advances in XAI and their aligned human studies. We propose four novel technical contributions — a multimodal modular architecture (MultiModN), an interpretable mixture-of-experts model (InterpretCC), adversarial training for explainer stability, and a theory-driven LLM-XAI framework (iLLuMinaTE) — evaluated with professors, teachers, learning scientists, and university students. Awarded the Patrick Denantes Memorial Prize for EPFL's best computer science thesis.
Thesis Pre-Print News
HEXED
2025EDM Workshop

2nd Human-Centric XAI in Education Workshop

Vinitra Swamy, Jakub Kuzilek, Juan D. Pinto, Luc Paquette, Tanja Käser, Qianhui Sophie Liu, Lea Cohausz
Co-organizer of the second edition of the HEXED workshop at the International Conference on Educational Data Mining, convening the community working on human-centric explainable AI for education.
EDM
Bloom
2025AIED

BloomTutor: Retrieval Augmentation for Bloom's Taxonomy Question Generation

Yannis Laaroussi, Vinitra Swamy, Paola Mejia-Domenzain, Adrien Vauthey, Aybars Yazici, Maxime Perrot, Tanja Käser
We present BloomTutor, an Intelligent Tutoring System that integrates Retrieval-Augmented Generation with Bloom's Taxonomy to support learners in exploring, revising, or querying topics within a course. The system retrieves course materials, segments them into chunks, and generates questions aligned with cognitive levels; learner responses trigger real-time difficulty adjustments toward higher-order thinking or reinforcement of foundational concepts. Unlike traditional ITS platforms that rely on static question repositories, BloomTutor generates context-specific questions dynamically for broader coverage and responsiveness to learner needs.
Paper PDF
AI or 🧑?
2024ECTEL🏆 Best Paper

AI or Human? Student Feedback Perceptions

Tanya Nazaretsky, Paola Mejia-Domenzain, Vinitra Swamy, Jibril Frej, Tanja Käser
We investigate how the identity of the feedback provider affects students' perception, comparing AI-generated and human-created feedback. Students evaluate feedback in authentic educational settings both before and after the provider's identity is disclosed. Our study with 457 students across diverse programs reveals that the ability to differentiate depends on the task; disclosing identity leads to a greater preference for human-created feedback and a decreased evaluation of AI-generated feedback. Students who failed to identify the provider correctly tended to rate AI feedback higher, whereas those who succeeded preferred human feedback.
Paper Pre-Print News
3C
2024AIED🏆 Best LBR

Interpret3C: Interpretable Student Clustering

Isadora Salles, Paola Mejia-Domenzain*, Vinitra Swamy*, Julian Blackwell, Tanja Käser
Interpret3C (Interpretable Conditional Computation Clustering) is a novel clustering pipeline that incorporates interpretable neural networks in an unsupervised learning context. It leverages adaptive gating to select features per student, then clusters using the most relevant features for each, enhancing cluster relevance and interpretability. We use Interpret3C to analyze behavioral clusters in a MOOC with over 5,000 students, offering a scalable clustering methodology and a case study that respects individual student differences.
Pre-Print Code
MCQ
2024EDM

Student Answer Forecasting

Elena Grazia Gado, Tommaso Martorella, Luca Zunino, Paola Mejia-Domenzain, Vinitra Swamy, Jibril Frej, Tanja Käser
Recent research on intelligent tutoring systems has focused on answer correctness rather than performance on specific answer choices, limiting insight into student thought processes and misconceptions. We present MCQStudentBert, an answer forecasting model that leverages LLMs to integrate contextual understanding of students' answering history along with the text of questions and answers. Practitioners can extend the model to new answer choices without retraining. We compare MLP, LSTM, BERT, and Mistral 7B architectures on a language learning MCQ dataset from an ITS with over 10,000 students.
Pre-Print Code
MultiModN modular multimodal architecture
2023NeurIPS

MultiModN: Multimodal, Multi-Task, Interpretable Modular Networks

Vinitra Swamy*, Malika Satayeva*, Jibril Frej, Thierry Bossy, Thijs Vogels, Martin Jaggi, Tanja Käser, Mary-Anne Hartley
Predicting multiple real-world tasks in a single model requires understanding across modalities. MultiModN is a multimodal, modular network that fuses latent representations in a sequence of any number, combination, or type of modality, while providing granular real-time predictive feedback on any number of tasks. Unlike parallel fusion, MultiModN is inherently interpretable, resistant to biased missingness, and can be composed of any pre-trained module. Evaluated on ten real-world tasks spanning text, images, tabular and time series, it matches parallel-fusion performance without the catastrophic failure under biased missingness that dooms conventional models.
Pre-Print Code News Poster
70B
2023Nature Preprint

MEDITRON-70B: Scaling Medical Pretraining for LLMs

EPFL LLM Team, International Committee of the Red Cross
MEDITRON is a suite of open-source LLMs with 7B and 70B parameters adapted to the medical domain, built on Llama-2 and extending pretraining on a comprehensively curated medical corpus. Evaluations against four major medical benchmarks show significant performance gains over several state-of-the-art baselines, with a 6% absolute improvement over the best public baseline and 3% over the strongest finetuned Llama-2. MEDITRON-70B outperforms GPT-3.5 and Med-PaLM and is within 5% of GPT-4 and Med-PaLM-2.
Pre-Print Journal Code Models News
Bias
2023EMNLP Findings

Unraveling Downstream Gender Bias from LLMs

Farnaz Kohankhaki, D. B. Emerson, Jacob-Junqi Tian, Laleh Seyyed-Kalantari, Faiza Khan Khattak (with Vinitra Swamy collaboration on the AI writing support study)
We investigate how bias transfers through an AI writing support pipeline in a large-scale user study with 231 students writing business case peer reviews in German. Students are split into five groups with different levels of writing support (traditional ML suggestions, no assistance, and finetuned GPT2, GPT-3, GPT-3.5). Using GenBit, WEAT, and SEAT, we evaluate gender bias in model embeddings, in generated suggestions, and in the reviews students actually wrote. We find no significant difference in gender bias between groups with and without LLM suggestions — our research is therefore optimistic about the use of AI writing support in the classroom, showcasing a context where bias in LLMs does not transfer to students' responses.
Paper
Trust?
2023LAK🏆 Hon. Mention

Trusting the Explainers

Vinitra Swamy, Sijia Du, Mirko Marras, Tanja Käser
We validate explainers for student success prediction models through the eyes of the educators they are meant to serve. Comparing five popular XAI methods across two MOOC datasets, we find systematic disagreement between explainers on which features matter — then run a study with 26 professors whose ranked preferences reveal which explanations educators actually trust and can act on.
Pre-Print Paper Code
〜〜
2023AAAI

RIPPLE: Concept-Based Interpretation for Raw Time Series

Mohammad Asadi, Vinitra Swamy, Jibril Frej, Julien Vignoud, Mirko Marras, Tanja Käser
Time series are ubiquitous but notoriously hard to interpret at the level of human-meaningful concepts. RIPPLE brings concept-based interpretability to raw multivariate time series, letting practitioners reason about model behaviour in terms of domain concepts rather than raw signal values.
Pre-Print Paper
9,165
2022COLING

Bias at a Second Glance

Vinitra Swamy, Malika Satayeva, Thierry Bossy, Tanja Käser, et al.
We analyze bias across text and through multiple architectures on a corpus of 9,165 German peer-reviews collected from university students over five years, including helpfulness, quality, and critical aspect ratings plus demographic attributes. We conduct WEAT analyses on the corpus, on common pre-trained German language models (T5, BERT, GPT-2) and GloVe embeddings, and on those models after finetuning. In contrast to our expectations, the corpus itself reveals few biases — but the pre-trained German language models find substantial conceptual, racial, and gender bias and shift significantly along conceptual and racial axes during fine-tuning.
Pre-Print Paper
XAI
2022EDM

Evaluating the Explainers

Vinitra Swamy, Bahar Radmehr, Natasa Krco, Mirko Marras, Tanja Käser
A systematic comparison of five popular post-hoc explainability methods applied to the same student success prediction models, revealing pervasive disagreement about which features drive a prediction — a foundational result motivating the shift towards interpretable-by-design architectures.
Pre-Print Paper News
Meta
2022Learning@Scale

Meta Transfer Learning for Early Success Prediction

Vinitra Swamy, Mirko Marras, Tanja Käser
Early prediction of student success is hardest exactly when it matters most — in the first weeks of a course, with little data. We use meta transfer learning to transfer knowledge across MOOCs, improving early-weeks prediction for courses where behavioural history is still sparse.
Pre-Print Paper Code
KG
2021NeurIPS XAI

Interpreting Language Models Through Knowledge Graph Extraction

Vinitra Swamy, Angelika Romanou, Martin Jaggi
What does a language model learn, and when? We extract knowledge graphs from BERT-family models at successive points during pretraining, using the evolving graph structure as a lens on what factual knowledge the model acquires and in what order.
Pre-Print Poster Code
ONNX
2020Microsoft AI

ONNX: Open Neural Network eXchange

Vinitra Swamy — Microsoft AI Frameworks
Lead engineer on ONNX, the industry standard for AI model interoperability, collaborating with 30+ tech companies including Nvidia, Intel, Amazon, and Meta. Contributed to the ONNX converters, the model zoo, and ONNX Runtime, and presented research talks on model operationalization and acceleration at Microsoft MLADS and the UW eScience Institute.
Website Code Model Zoo
HXL
2019KDD

ML for Humanitarian Data: Tag Prediction with HXL

Vinitra Swamy, Sara Du, Megha Srivastava, Joseph M. Hellerstein
Humanitarian datasets are only useful when they are findable and comparable. We build ML models to automatically predict Humanitarian eXchange Language (HXL) tags for humanitarian data columns, reducing the manual annotation burden that keeps crisis data siloed.
Pre-Print News
DKT
2018AIED

Deep Knowledge Tracing for Student Code Progression

Vinitra Swamy, Allen Guo, Samuel Lau, Wilton Wu, Madeline Wu, Zachary Pardos, David Culler
Applying deep knowledge tracing to sequences of student code submissions in an introductory programming course, modelling how understanding develops across successive attempts rather than treating each submission independently.
Paper
Data 8
2018MSc Thesis🏆 Berkeley Fellow

Pedagogy, Infrastructure & Analytics for Data Science at Scale

Vinitra Swamy — UC Berkeley RISELab
A detailed research report on autograding, analytics, and scaling JupyterHub infrastructure for the thousands of students taking Data 8 at UC Berkeley. Helped develop the data science software stack including JupyterHub, autograding with OkPy and Gradescope, and authentication at scale. Collaborated with Yuvi Panda, Ryan Lovett, Chris Holdgraf, and Gunjan Baid on a JupyterCon 2017 talk detailing the infrastructure.
Thesis Data 8
CSi²
2017IBM Research

CSi2: Idle Server Identification

Vinitra Swamy, Neeraj Asthana, Sai Zheng — IBM T.J. Watson Research Center
"Zombie" virtual machines in hybrid and private clouds waste millions of dollars of resources. CSi2 is an ensemble ML algorithm to detect VM inactivity and recommend a course of action (termination, snapshot). Projected to save IBM Research at least $3.2M with 95.12% recall and 88% F1, and integrated into the Watson Services Platform. Two patents filed.
IBM Research

Industry

Scholé AI

Co-founder and CEO

Building personalized learning at scale. We recently raised $3M in venture funding, and launched as #1 product of the day on ProductHunt. Check out our multi-agent lesson delivery!

2025 - Current

Microsoft AI

AI Software Engineer

Working on a framework for deep learning / ML framework interoperability (ONNX) alongside an ecosystem of converters, containers, and inference engines.

Lead of the inter-company ONNX Special Interest Group (SIG) for Model Zoo and Tutorials with Microsoft, Intel, Facebook, IBM, nVidia, RedHat, and other academic and industry collaborators.

Presented and represented Microsoft AI at several conferences: WIDS 2020, Microsoft //Build 2019, KDD 2019, Microsoft Research Faculty Summit 2019, UC Berkeley AI for Social Impact Conference 2018, Women in Cloud Summit 2018, RISECamp 2018

2018 - 2020

Berkeley Insitute for Data Science (BIDS), RISELab

Research Assistant

Worked on projects in AI + Systems with an application area of data science education. Project areas include JupyterHub architecture, custom deployments, OkPy autograding integration, Jupyter noteboook extensions, and D3.js / PlotLy visualizations for data science explorations of funding and enrollment data.

[BIDS] [RISELab]

2015 - 2018

IBM Research

Research Scientist Intern, Machine Learning

Worked on the CSi2 project as a Machine Learning Research Scientist intern on the Hybrid Cloud team. The CSi2 algorithm is an ensemble machine learning algorithm to detect inactivity of VMs as well as suggest a course of action (i.e. termination, snapshot). It is projected to save IBM Research at least $3.2 million dollars with 95.12% recall and 88% F1 score (>> industry standard) and is being implemented into the Watson Services Platform. Collaborated with Neeraj Asthana, Sai Zheng, Ivan D'ell Era, Aman Chanana. Presented an exit talk and filed 2 patents.

2017

LinkedIn

Software Engineering Intern

Interned at LinkedIn headquarters with the Growth Division's Search Engine Optimization (SEO) Team the summer before entering UC Berkeley. Worked on fullstack testing infrastructure for the public profile pages, as well as a Hadoop project; outside of assigned work, helped plan LinkedIn’s DevelopHER Hackathon and worked on several Market Research/User Experience Design initiatives.

2015

Google

Intern, Made w/ Code Ambassador

Spent a summer learning computer science fundamentals and shadowing engineers through the CAPE high school internship program at Google Headquarters in Mountain View, CA. Chosen as a Google Ambassador for Computer Science following the experience. Worked with Google, Salesforce, and AT&T to introduce coding to over 15,000 girls across California with the Made w/ Code Initiative.

2011

Education

École Polytechnique Fédérale de Lausanne

PhD in Computer Science
  • President of EPFL PhDs in Computer Science (EPIC)
  • Advised by Prof. Tanja Käser at the ML4ED Lab
    and Prof. Martin Jaggi at the MLO Lab
  • EDIC Computer Science Fellowship Recipient
  • EPFL IC Distinguished Service Award (3-time Recipient)
  • Elected CS PhD Representative (EDIC Committee under Prof. Ed Bugnion)
2020 - 2025

University of California, Berkeley

Master's in Electrical Engineering and Computer Science
  • President of Computer Science Honor Society (UPE)
  • ✨🗞️Head Graduate Student Instructor of Data 8 (Foundations of Data Science)
  • Research Assistant, Graduate Opportunity Fellow at RISELab
  • Advisor: Dean of Data Sciences, David Culler
2017 - 2018

University of California, Berkeley

Bachelor's in Computer Science
  • EECS Award of Excellence in Undergraduate Teaching and Leadership
  • UC Berkeley Alumni Leadership Scholar
  • Graduated 2 years early
2015 - 2017

Teaching Experience

Teaching Faculty for Harvard DSI AI Intensives

Harvard University
  • Teaching Faculty to 1000s of professionals around the world through Harvard Data Science Initiative's GenAI, Agentic AI, and AI Leadership Intensives. Forbes recently recommended the course as the top way to learn agents in 2026! Early Career Board for the Harvard Data Science Review.

[Course Announcement] [Agentic AI Intensives]

2025 - Current

Guest Lecturer, TA for Machine Learning for Behavioral Data (CS 421)

EPFL

2020 - 2024

Adjunct Faculty for CSE/STAT 416: Introduction to Machine Learning

University of Washington, Seattle
  • Lecturing to 100+ upper-division undergraduate and graduate students on a practical introduction to machine learning. Modules include regression, classification, clustering, retrieval, recommender systems, and deep learning, with a focus on an intuitive understanding grounded in real-world applications.

[CSE 416 Website]

Summer 2020

Lecturer for Data 8: Foundations of Data Science

UC Berkeley
  • ✨🗞️ Lecturing to 250+ undergraduate students on fundamentals of statistical inference, computer programming, and inferential thinking.

[Data 8 Website] [Course Offering] [Course Materials / Code]

Summer 2018

Head TA for Data 8: Foundations of Data Science

UC Berkeley
  • ✨🗞️ TA / Head Graduate Student Instructor (GSI) of Data 8 for 4 semesters, responsible for management of 1000+ undergraduates, 40 GSIs, 30 tutors, and 100+ lab assistants each semester.
  • ✨🗞️ Helped create data science curriculum material for lecture and domain-specific seminar courses.
  • ✨🗞️ In charge of developing JupyterHub infrastructure for 1500+ active users (with Jupyter Servers with Docker/Kubernetes backend on top of various cloud providers including Google Cloud, Azure, and AWS).
2016 - 2018

Organizing Team

HEXED Workshop Organizer @ EDM 2024
WiML Program Chair @ ICML 2022
FATED Workshop Organizer @ EDM 2022

Reviewer / Program Committee

AIED Program Committee 2023, 2024
AIED 2021*, 2022* (Subreviewer for Tanja Käser)
EMNLP BlackBoxNLP 2021, 2022, 2023
NeurIPS 2024
NeurIPS GenBench 2022, 2023, GAIED 2023, and XAI Workshops 2023, 2024
EACL 2022
EDM Program Committee 2023, 2034
Journal of Educational Data Mining (JEDM) 2022, 2023
LAK 2022*, 2023* (Subreviewer for Tanja Käser)
Editor for Springer Series on Big Data Management (Educational Data Science)

Working Groups

Fairness Working Group @ EDM 2022
WiML Workshop Team @ NeurIPS 2021
Lead of the 2020 ONNX SIG for Models and Tutorials

2023

Awards


Speaking Engagements

  • Spring 2026: Speaker at the ASU+GSV Summit in San Diego, presenting Scholé AI as part of the GSV Cup 50
  • Spring 2026: Speaker at ACE Ventures Investor Day, presenting Scholé AI's agentic learning engine
  • Fall 2025: Speaker at EdTech Week in New York City on AI-native workforce learning
  • Fall 2024: Talk at Rising Stars in Data Science Workshop at UC San Diego on the "Future of human-centric eXplainable AI"
  • Summer 2024: Invited Talk at Microsoft Research Cambridge in Cambridge, UK on evaluating post-hoc explainers, InterpretCC, MultiModN, and iLLuMinaTE
  • Summer 2024: Speaker at the Data Makers Fest 2024 in Porto, Portugal on the future of Explainable AI
  • Summer 2024: Speaker at the LauzHack Deep Learning Bootcamp on Advanced Topics: Explainable AI
  • Spring 2023: Speaker at the SMART-AI Workshop for the WHO on Interpretable AI
  • Spring 2023: Speaker at the Applied XAI track of Applied Machine Learning Days 2024
  • Spring 2023: Speaker at EDIC Open House for the ML4ED Laboratory
  • Spring 2023: Instructor at the BeLEARN center for the JDPLS program (ETH Zurich, EPFL) on Student Modeling
  • Fall 2023: Speaker at Red Cross LLM Day at the ICRC Headquarters on Evaluating and Interpreting LLMs
  • Fall 2023: Speaker at AWS Research Day on Personalized, Trustworthy Human-Centric Computing: AI for Education
  • Summer 2022: Speaker at Oxford ML "Un-Workshop" Series on Evaluating Explainable AI
  • Summer 2022: Opening Remarks at the FATED workshop at EDM 2022 (Durham, UK)
  • Spring 2022: Speaker at Women in Data Science (WIDS 2022) Silicon Valley: Explainable AI
  • Fall 2021: Spotlight Talk at NeurIPS Inaugural eXplainable AI for Debugging and Diagnosis Workshop
  • Fall 2021: Presenter at the Tamil Internet Conference (INFITT) on "TamilBERT: Natural Language Modeling for Tamil"
  • Spring 2021: Presenter at the EDIC Orientation for PhDs, EPFL
  • Spring 2021: UC Berkeley Data Science Alumni Panel (Data 8)
  • Fall 2020: Featured Guest on the Tech Gals Podcast (Episode 3)
  • Fall 2020: Speaker at the ONNX Workshop
  • Spring 2020: Speaker at Women in Data Science Conference (WIDS 2020) Silicon Valley: Interoperable AI (ONNX)
  • Spring 2020: Speaker at the Linux Foundation (LF) AI Day
  • Fall 2019: Presenter at Microsoft Bay Area AI Meetup
  • Summer 2019: Guest on the Microsoft AI Show (Channel 9)
  • Spring 2019: Speaker at Microsoft Machine Learning and Data Science Conference (MLADS) (Redmond)
  • Summer 2018: Presenter at Artificial Intelligence in Education 2018 (London)
  • Summer 2018: Speaker at UC Berkeley's Data Science Undergraduate Pedagogy and Practice Workshop (Berkeley)
  • Spring 2018: Brilliance of Berkeley Panelist on the "Transformative Powers of Data" (LA '18)
  • Fall 2017: Opening Panelist at SalesForce Dreamforce Conference (SF)
  • Summer 2017: Speaker at JupyterCon (NYC)
  • Spring 2017: Presenter at Berkeley Institute for Data Science Research Showcase (Berkeley)
  • Fall 2016: Panelist at SF BusinessWeek Conference (SF)
  • Summer 2016: Conference organizing team at Algorithms for Modern Massive Data Sets (MMDS) (Berkeley)

News