Devanshu Shah

Research Assistant with Prof. Max Simchowitz and Prof. Aviral Kumar at CMU

Handle

@devanshu

Role

Researcher

Company

Carnegie Mellon University (verified: @cs.cmu.edu)

Location

Pittsburgh, PA

About

I research reinforcement learning, flow-matching models, continual learning for LLMs, and efficient machine learning systems. I build methods that improve sample efficiency and training speed, including CUDA Graphs and in-place state restoration.

Experience

• Researcher at Carnegie Mellon University

Education

• Carnegie Mellon University
  Master of Science, Machine Learning
  2025 to 2026
  Courses: Adv. Intro. to ML, Deep Reinforcement Learning, Prob. & Math. Statistics, Machine Learning Systems, Diffusion and Flow Matching.
• Birla Institute of Technology and Science, Pilani
  Bachelor of Engineering, Computer Science; Minor: Data Science
  2018 to 2022
  CGPA: 9.09/10, K.K. Birla Goa Campus.

Projects

• Adaptive Expert Parallelism in MoEs
  Mar 2026 to Apr 2026
  Formulated a dynamic expert-placement optimizer achieving a 1.165x MoE-layer speedup for OLMoE at equal memory. Modeled GroupGEMM compute and all-to-all communication on A100s and optimized expert placement using ILP.
• COVID-19 Forecaster
  Mar 2021 to Apr 2021
  Developed methods to forecast COVID-19 cases and achieved an 11.53% MAPE, a 54% improvement over baseline. Built LSTM and N-BEATS models in PyTorch and used meta-learning to improve sample efficiency.

Skills

• Machine Learning
• Reinforcement Learning
• flow-matching models
• CUDA Graphs
• LLMs
• vector database
• OpenSearch
• Bazel
• PyMC3
• scikit-learn
• PyTorch
• ILP

This profile as JSON