Achille Nazaret

Ph.D. Student in Computer Science, Columbia University

I am a Ph.D. student in computer science advised by Prof. David Blei and Prof. Elham Azizi.

I develop mathematical models to describe and understand the world. I specialize in probabilistic and generative models, focusing on (i) scaling them to large datasets and (ii) interpreting what they learned. For example, I train generative models on Apple Watch sensor data to understand subjects' health.

I orient my research towards practical impact, prioritizing simplicity and usability over method complexity. I value teamwork, mathematical rigor, clean code implementation, and clear scientific communication.

Resume

Find a PDF version here.

Education

Columbia University

New York, NY

Ph.D. in computer science.

Jan 2021 - May 2025

M.S. in computer science.

Aug 2019 - Dec 2020

École Polytechnique

Palaiseau, France

M.S. in applied mathematics and computer science.

Aug 2018 - Apr 2019

B.S. in applied mathematics and computer science.

Aug 2016 - Jun 2018

École Spéciale Militaire de Saint-Cyr

Coëtquidan, France

Accelerated track to the rank of army officer (second lieutenant).

Aug 2016 - Nov 2016

Lycée Privé Sainte-Geneviève

Versailles, France

Prépa in pure mathematics, physics, and computer science.

Aug 2014 - Jul 2016

Experience

Apple Health AI

New York, NY

Machine learning research scientist (part-time, alongside Ph.D.)

Feb 2024 - Dec 2024
- Foundation models of time series from wearables to understand user health and fitness levels.
Apple Health AI

New York, NY

Machine learning research scientist (intern)

Jan 2022 - Aug 2022
- Estimated the causal impact of the Watch's notifications on user behavior with novel causal estimators.
- It increases standing rates by 40%.
Apple Health AI

New York, NY

Machine learning research scientist (intern)

May 2021 - Aug 2021
- Identified and designed new health biomarkers from the sensor data of Apple devices.
Palantir Technologies

(remote) San Francisco, CA

Forward deployed software engineer (intern)

Jun 2020 - Aug 2020
- Scoped, prototyped, and deployed data-driven algorithms to reduce costs for a US healthcare insurer.
Yosef Lab, University of California, Berkeley

Berkeley, CA

Research assistant

Apr 2019 - Aug 2019
- Developed an open-source Python package for single-cell data analysis: scvi-tools (1.3k+ GitHub stars).
- Designed generative models to impute unobserved genes in spatial genomics using sc-RNA data.
Akwa Group

(remote) Casablanca, Morocco

Machine learning consultant (alongside M.S.)

Sep 2018 - May 2019
- Designed signals and models to predict the performance of new gas stations -- surpassed human experts by 25%.
IMC Trading

Amsterdam, Netherlands

Software engineer (intern)

Jun 2018 - Sep 2018
- Distributed model training pipelines on a cluster for faster overnight training (HFT, futures).
Bernardaud

Paris, France

Operations research consultant (alongside B.S.)

Feb 2018 - Jun 2018
- Designed algorithms to find optimal production processes under factory constraints.
- Created a user-friendly full-stack website connecting my algorithms to the databases.
Ministry of Defense

Paris, France

Junior data scientist (intern)

Nov 2016 - Apr 2017
- Developed graph-mining and NLP models for social network analysis to produce intelligence

Publications

You can find my full list of publications on Google Scholar.

Here are my main first-author publications (* indicates co-first):

N. Beltran-Velez*, A. A. Grande*, A. Nazaret*, A. Kucukelbir, D. Blei. Treeffuser: Probabilistic Predictions via Conditional Diffusions with Gradient-Boosted Trees

NeurIPS 2024
C. Shi*, N. Beltran-Velez*, A. Nazaret*, C. Zheng*, A. Garriga-Alonso, A. Jesson, M. Makar, D. Blei. Hypothesis testing the circuit hypothesis in LLMs

NeurIPS 2024
An earlier version of the work above was presented at the ICML 2024 Workshop on Mechanistic Interpretability.

ICML 2024 Workshop Best paper (3rd)
A. Nazaret*, J. Hong*, E. Azizi, D. Blei. Stable Differentiable Causal Discovery

ICML 2024
A. Nazaret, D. Blei. Extremely Greedy Equivalence Search

UAI 2024
J. Fan*, A. Nazaret*, J. Fan*, E. Azizi. A thousand and one tumors: the promise of AI for cancer biology

Nature Methods 2024
A. Nazaret*, J. Fan*, V. Lavallée, D. Pe'er, E. Azizi. Deep generative modeling for mapping derailed trajectories in Acute Myeloid Leukemia

bioRxiv preprint 2024
A. Nazaret, C. Shi, D. Blei. On the Misspecification of Linear Assumptions in Synthetic Control

AISTATS 2024
S. He*, Y. Jin*, A. Nazaret*, L. Shi, X. Chen, S. Rampersaud, E. Azizi et al. Starfysh integrates spatial transcriptomic and histologic data to reveal heterogeneous tumor–immune hubs

Nature Biotechnology 2024
A. Nazaret, S. Tonekaboni, G. Darnell, S. Ren, G. Sapiro, A. Miller. Modeling Heart Rate Response to Exercise with Wearables Data

Nature Digital Medicine 2023
A. Nazaret, G. Sapiro. A large-scale observational study of the causal effects of a behavioral health nudge

Science Advances 2023
A. Nazaret, J. Hong. BetterBoost - Inference of Gene Regulatory Networks with Perturbation Data - GSK.ai CausalBench Challenge. Machine Learning for Drug Discovery Workshop.

ICLR 2023 Challenge First place
A. Nazaret, J. Fan, D. Pe'er, E. Azizi. Probabilistic basis decomposition for characterizing temporal dynamics of gene expression - Workshop on Computational Biology.

ICML Workshop 2022 Best poster award
A. Nazaret, D. Blei. Variational Inference for Infinitely Deep Neural Networks

ICML 2022
K. Choromanski*, D. Cheikhi*, J. Davis*, V. Likhosherstov*, A. Nazaret*, A. Bahamou*, X. Song*, M. Akarte, J. Parker-Holder, J. Bergquist, Y. Gao, A. Pacchiano, T. Sarlos, A. Weller, V. Sindhwani.. Stochastic flows and geometric optimization on the orthogonal group

ICML 2020
R. Lopez*, A. Nazaret*, M. Langevin, J. Samaran, J. Regier, M. I Jordan, N. Yosef. A joint model of unpaired data from scRNA-seq and spatial transcriptomics for imputing missing gene expression measurements - Workshop on Computational Biology.

ICML Workshop 2019 Best poster award

Here are other publications:

M. Chevalley, J. Sackett-Sanders, Y. Roohani, P. Notin, A. Bakulin, D. Brzezinski, K. Deng, Y. Guan, J. Hong, M. Ibrahim, W. Kotlowski, M. Kowiel, P. Misiakos, A. Nazaret, M. Püschel, et al. The CausalBench challenge: A machine learning contest for gene network inference from single-cell perturbation data

arXiv preprint 2023
A. Gayoso, R. Lopez, G. Xing, P. Boyeau, J. Hong, K. Wu, M. Jayasuriya, E. Mehlman, M. Langevin, Y. Liu, J. Samaran, G. Misrachi, A. Nazaret, O. Clivio, C. Xu, T. Ashuach, M. Gabitto, M. Lotfollahi, V. Svensson, E. Beltrame, V. Kleshchevnikov, C. Talavera-López, L. Pachter, F. J Theis, A. Streets, M. I Jordan, J. Regier, N. Yosef.. A Python library for probabilistic analysis of single-cell omics data

Nature Biotechnology 2022

Open source software

Treeffuser An easy-to-use package for probabilistic prediction on tabular data with tree-based diffusion models.
SDCD A method for inferring causal graphs from labeled interventional data.
scvi-tools A library for analyzing single-cell data with deep generative models.