Chandler Squires
Chandler Squires

Postdoctoral Research Associate

CMU

Since September 2024, I have been a Postdoctoral Research Associate in the Machine Learning Department at CMU, supervised by Pradeep Ravikumar. Previously, I was a PhD Student at MIT, where I was advised by Caroline Uhler and David Sontag.

My research focuses on the foundations and practical applications of Neurosymbolic AI, broadly construed as a discipline which integrate symbolic and nonsymbolic approaches for representation and computation. Within this space, I focus on symbolic representation learning and its downstream applications, including model steering, unlearning, and representation alignment.

My research agenda is heavily influenced by pragmatist philosophy and related principles, e.g. Vapnik’s famous quote: “When solving a problem of interest, do not solve a more general problem as an intermediate step”. In my PhD thesis , I summarize this research philosophy as Pragmatic Data Science, which emphasizes a decision-centric viewpoint of machine learning and statistics. In line with this view, my PhD research focused on various problems in causality, including causal structure learning, causal representation learning, and causal effect estimation.

Favorite Flavors of Theory: My theoretical work uses tools from high-dimensional statistics, identifiability theory, semiparametric efficiency theory, Bayesian experimental design, and combinatorial optimization.

Favorite Applications: On the application side, I am broadly interested in AI4Science for AI4Operations, and the unique challenges encountered when bringing AI into scientific and engineering domains. Right now, I am focused on applications of my work in drug discovery, cellular biology, and manufacturing.

(Last update: September 25th, 2025)


Teaching and Service. I developed and taught a seven-lecture course on causality for MIT’s 2023 January term, check out the lecture notes and recordings.

During my PhD, I was a member of the EECS Communication Lab, where I coached researchers on technical communication tasks such as paper writing, poster design, and oral presentation.

Software. During my PhD, I developed and maintained causaldag, a Python package for creating, manipulating, and learning causal graphical models.

Other materials. I try to keep an up-to-date repository of slides from my talks and posters from conferences and workshops, but these are ~1 year out of date.


Interests
  • Causality
  • Representation Learning
  • Experimental Design
  • Neurosymbolic AI
  • Computational Biology/Genomics
Education
  • PhD in Electrical Engineering and Computer Science, 2024

    Massachusetts Institute of Technology

  • MEng in Electrical Engineering and Computer Science, 2019

    Massachusetts Institute of Technology

  • BSc in Electrical Engineering and Computer Science, 2018

    Massachusetts Institute of Technology

Featured Publications
Recent Publications
(2025). Knowledge-Enriched Machine Learning for Tabular Data. NeuS 2025.
(2025). Probably Approximately Correct High-dimensional Causal Effect Estimation Given a Valid Adjustment Set. CLeaR 2025.
(2025). The Landscape of Causal Discovery Data: Grounding Causal Discovery in Real-World Applications. CLeaR 2025.
(2025). Synthetic Potential Outcomes for Mixtures of Treatment Effects}. AISTATS 2025.