Sam Stites

My name is Sam Stites – I'm a Ph.D. student at the Northeastern PRL studying Bayesian inference and the semantics of probabilistic programming languages (PPLs). Before joining the PRL in 2022 I was part of the data@NU lab researching deep probabilistic programming. Prior to starting my Ph.D. program I was in industry for a 7-year stretch. I wrote Hasktorch, the Haskell bindings to PyTorch.

In my ongoing research, I study the interoperation between different probabilistic languages which live in different semantic domains. The result of this is a flexible inference strategy that can overcome shortcomings found in each individual language, so long as we maintain certain safety properties. More broadly, I'm interested in semantics, type theory, and formal verification; I often find myself typing away in an Agda file just to make sense of things.

Contact: The best way to contact me is by email. Apply ROT13 to the following: [email protected]. I'm also reachable as .stites on discord.

#Code

I work on a rust compiler and formalization that prototypes a multi-language approaches to probabilistic inference.

A small project from my past which I actively maintain is redirect-to-abstract (github). If this piques your interest and you want to use it, feel free to reach out!

Projects I no longer work on, but are still going strong:

Unmaintained projects:

  • reinforce reinforcement learning in Haskell.
  • CSSR (v2). Causal State Splitting Reconstruction of recursive hidden Markov models. I coded this up with Cosma Shalizi before starting my Ph.D. studies and it is currently in an unpublished state. If this work is relevant to you, please contact Cosma directly and cc me.

Feel free to reach out if want to talk about stenography, Agda, or parenting in grad school.

#Teaching

Teaching Assistant positions include:

In addition, I helped develop and co-teach a "Mechanizing Metatheory" module for Chris Martens' CS7800 Intensive Principles of Programming Languages in Fall 2024.

#Publications and Talks

My latest submission has been conditionally accepted for publication in 20251! Published works include:

Sam Stites, Heiko Zimmermann, Hao Wu, Eli Sennesh, Jan-Willem van de Meent Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence, PMLR 161:1056-1066, 2021. (abstract, pdf, code)

Invited talks include:

  • LAFI 2024: A Multi-language Approach to Probabilistic Program Inference. Sam Stites, Steven Holtzen
  • NPFL 2018: Hasktorch: A Comprehensive Haskell Library for Differentiable Functional Programming. Sam Stites, Austin Huang

#Service

  • SIGPLAN-M Mentor (2023-)
  • PLDI 2023 Student Volunteer
  • AISTATS 2022 Reviewer
  • Northeastern 2022 Faculty Admissions Volunteer
  • Northeastern 2021 Ph.D. Review Committee Volunteer

#Webring

Footnotes


1

I'm not really sure how much I can state before de-anonymizing myself, but this is my first big PL submission and was a sizable amount of work – I'm jazzed!