Kurt Shuster
Member Of Technical Staff at Thinking Machines Lab
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolKurt Shuster is an AI research engineer and Member of Technical Staff with 11 years of experience building production-grade machine learning infrastructure and conversational AI systems. He holds an MEng in Computer Science from Cornell and has worked across industry-leading teams at Facebook AI, Character.AI, Google DeepMind, and now as a founding engineer at Thinking Machines Lab. Kurt combines core research instincts with pragmatic backend and DevOps expertise—evidenced by substantial contributions to the ParlAI dialogue framework, including automated Slurm parameter sweeps and core teacher-module improvements. He excels at turning experimental models into reproducible, scalable pipelines and has a track record of shipping infrastructure that supports large-scale dialogue research. Based in New York, he brings both research rigor and hands-on systems craftsmanship to applied AI problems.
11 years of coding experience
8 years of employment as a software developer
Master of Engineering - MEng Computer Science, Master of Engineering - MEng Computer Science at Cornell UniversityBachelors Computer Science, Bachelors Computer Science at Cornell University College of Arts and Sciences
Github Skills (12)
automation10
parameter-tuning10
bash10
automations10
hpc10
slurm10
optimisation10
python10
optimization10
ml-deployment9
machine-learning9
continuous-deployment9
Programming languages (2)
Github contributions (5)
facebookresearch/ParlAI
Oct 2017 - Dec 2022
A framework for training and evaluating AI models on a variety of openly available dialogue datasets.
Role in this project:
Backend & DevOps Engineer
Backend & DevOps EngineerContributions:5 releases, 536 reviews, 696 commits in 5 years 3 months
Contributions summary:Kurt contributed significantly to the backend infrastructure by adding and modifying scripts for automated parameter sweeps using Slurm, which is evident from the shell script additions within the "projects" folder. They demonstrated expertise in automating machine learning model training, parameter tuning, and deployment infrastructure, as evidenced by the scripts for running parameter sweeps for bAbI tasks using a Memory Network model. Additionally, the user modified core modules within the library, such as in `FbDialogTeacher`, demonstrating a focus on the core functionality of the ParlAI framework. The creation of a generalized save root further indicates work in developing the infrastructure for the framework.
nlpdeep-learningdatasetmachine-learningtraining
facebookresearch/talkthewalk
Apr 2018 - Aug 2019
This repository provides code for reproducing experiments of the paper Talk The Walk: Navigating New York City Through Grounded Dialogue by Harm de Vries, Kurt Shuster, Dhruv Batra, Devi Parikh, Jason Weston, and Douwe Kiela.
Contributions:22 commits, 4 PRs, 4 pushes in 1 year 4 months
nlpnew-york-citydeviwalkdialogue
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.