David Rémy Bellamy

Since 2012, I went from being a biochemist to an AI researcher and bring a rare interdisciplinary skillset to the field. I am currently a founding AI researcher at Lila Sciences, a startup originating from Flagship Pioneering, where Moderna was founded. At work, I am developing post-training techniques and evaluations to teach an LLM agent how to conduct scientific research autonomously. My prior work has been featured in venues such as NeurIPS[1] [2], ICML [3], Machine Learning for Health (ML4H) [4], the European Journal of Epidemiology[5], and Nature Medicine[6] as well as an mRNA patent with the USPTO. I am an experienced software developer and obsess over high-quality code and research infrastructure to accelerate experimentation and learning.

I think that AI is going to transform the world more significantly than computers. I am excited to be a part of this transformation.

News

12/2024: We publicly launched our company Lila Sciences!
12/2024: Labrador was awarded the 2024 Machine Learning for Health (ML4H) Best Paper Award.
11/2024: Labrador was awarded an oral spotlight presentation at ML4H 2024.
10/2024: DAG-aware Transformer accepted at NeurIPS 2024 workshop on Causal Representation Learning.
10/2023: GPT-4 evaluation on the US Anaesthesiology Board Exam accepted by the British Journal of Anaesthesia.
06/2023: Conformal prediction with LLMs accepted at ICML 2023 workshop on Neural Conversational AI.
05/2023: Joined Lila Sciences as employee #1 and founding AI researcher.
05/2023: I defended my PhD on the 2nd of May. My thesis is here. Thanks to my examination committee Andrew Beam, Tianxi Cai and Leo Celi.
09/2022: Deep learning for proximal inference paper accepted at NeurIPS 2022 main conference.
07/2022: Began consulting with Artera.ai on a causal machine learning project for precision treatment decision-making.
06/2022: Structural characterization of shortcut features accepted by the European Journal of Epidemiology.
03/2022: Began teaching Harvard’s Deep Learning course at the Medical School.
02/2022: Placed 2nd out of 60 teams in the Lab for Innovation Science at Harvard datathon.
01/2022: Attended the Causality Boot Camp at The Simons Institute for the Theory of Computing.
05/2022: I completed my Masters of Science in Biostatistics at Harvard University.
01/2022: I was awarded the Harvard Department of Epidemiology’s UCB Pharma Fellowship.
09/2021: I was selected as a founding member of the Harvard CAUSALab by Miguel Hernan and James Robins.
06/2021: Began teaching the full-time graduate summer foundations course in biostatistics.
10/2020: I published a systematic review of neural net performance on tabular healthcare data.
08/2020: Analysis of nonprofit vs. for-profit charity care spending accepted at the Journal of General Internal Medicine.
06/2020: I joined Andrew Beam’s lab at Harvard officially marking my transition to machine learning.
08/2018: Started a second PhD in Epidemiology at the Harvard School of Public Health to study algorithms for digital phenotyping with Elise Robinson.
01/2018: Left my PhD at Harvard Medical School to pursue a more ML-focused path.
08/2016: Started my PhD in Biological & Biomedical Sciences at Harvard Medical School to study transgenerational epigenetic inheritance in C. elegans with Eric Greer.
01/2016: Began teaching the University of Ottawa’s core Biochemistry course.
05/2015: Read Hinton, Bengio and LeCun’s 2015 Nature paper “Deep learning”. Curiosity for AI began.
01/2015: Paper on oncolytic viruses and the tumor microenvironment accepted at Nature Medicine.
09/2014: Joined Harvard Medical School as a student researcher in Yang Shi’s lab.

Harvard PhD

In my PhD, I built a pre-training dataset, bespoke data tokenizer for a non-canonical modality and a continuous embedding layer to adapt the Transformer to risk prediction in the intensive care unit. I also developed a neural net estimator to solve an ill-posed integral equation for causal inference in the presence of unmeasured confounding.

Interests

I am a perpetual learner with a lot of curiosity. In my spare time, I do my own LLM research, read math books, solve programming and probability puzzles, learn new languages and read papers from fields like neuroscience, drug discovery and psychiatry. I also love the outdoors, traveling and watching movies!

Before AI

I got my start in the lab when I was 18 years old, developing the synthesis for a large organic ligand to coordinate Rhenium and lower the energy barrier for CO2 reduction. I proceeded to work in 6 different laboratories spanning the engineering of oncolytic viruses, the epigenetics of leukemia, mass spectrometry for peptidomics, and transgenerational epigenetic inheritance. I began a PhD in Molecular Biology at Harvard Medical School in 2016 simultaneously enrolling in Harvard’s Neuroscience PhD curriculum but by then was already so impacted by LeCun, Bengio and Hinton’s 2015 Nature paper “Deep learning” that I had to make a career change. Fast forward several years, I received my PhD in Epidemiology and Masters in Biostatistics at the Harvard School of Public Health in May 2023 under the supervision of Andrew Beam, Tianxi Cai and Leo Celi.

On the causality side, I was a member of the Harvard CAUSALab, where I was privileged to work with and learn from Miguel Hernán and James Robins. I have also consulted with Artera.ai on a causal machine learning project for precision treatment decision-making. Upon graduating, I joined my current company as a founding AI scientist.