Research Scientist – Deep Learning
Research Scientist – Deep Learning
Research Scientist – Deep Learning
ABOUT ASTRUS
ABOUT ASTRUS
ABOUT ASTRUS
📍 Location: Toronto or Waterloo, Canada
At Astrus, we are using AI to automate microchip design, starting with the biggest bottleneck, analog layout. Our mission is to radically improve global computation and empower chip designers to create the world's most advanced microchips with AI. Astrus is backed by top-tier VC firms: Khosla Ventures, HOF Capital, and 1517 Fund.
📍 Location: Toronto or Waterloo, Canada
At Astrus, we are using AI to automate microchip design, starting with the biggest bottleneck, analog layout. Our mission is to radically improve global computation and empower chip designers to create the world's most advanced microchips with AI. Astrus is backed by top-tier VC firms: Khosla Ventures, HOF Capital, and 1517 Fund.
ABOUT THE ROLE
ABOUT THE ROLE
ABOUT THE ROLE
We’re hiring a Research Scientist to work alongside our research team to push the boundaries of neural network architecture and generalization in simulated physical design.
You’ll take the lead on shaping our core deep learning stack — helping our models learn faster, scale better, and generalize across novel layout challenges. This role is research-heavy and exploratory: you’ll propose and validate new architectures, bring structure to unstructured design problems, and help shape the future of how AI builds physical compute.
You’ll collaborate closely with our reinforcement learning and infra teams, and operate in a highly technical environment with deep autonomy and strong ownership.
We’re looking for someone who thrives in ambiguous settings, moves fast, and cares deeply about elegant ideas and working systems.
We’re hiring a Research Scientist to work alongside our research team to push the boundaries of neural network architecture and generalization in simulated physical design.
You’ll take the lead on shaping our core deep learning stack — helping our models learn faster, scale better, and generalize across novel layout challenges. This role is research-heavy and exploratory: you’ll propose and validate new architectures, bring structure to unstructured design problems, and help shape the future of how AI builds physical compute.
You’ll collaborate closely with our reinforcement learning and infra teams, and operate in a highly technical environment with deep autonomy and strong ownership.
We’re looking for someone who thrives in ambiguous settings, moves fast, and cares deeply about elegant ideas and working systems.
WHAT YOU WILL DO
WHAT YOU WILL DO
WHAT YOU WILL DO
Design and experiment with novel neural network architectures tailored to the layout generation problem
Improve the generalization of our models across a wide range of simulated analog blocks
Propose and lead new research ideas — from hypothesis to implementation
Design targeted experiments to validate ideas and explore scaling behaviours
Collaborate with RL researchers and engineering leads to shape an end-to-end intelligent system
Help maintain a high research bar and push Astrus toward breakthrough capability
Design and experiment with novel neural network architectures tailored to the layout generation problem
Improve the generalization of our models across a wide range of simulated analog blocks
Propose and lead new research ideas — from hypothesis to implementation
Design targeted experiments to validate ideas and explore scaling behaviours
Collaborate with RL researchers and engineering leads to shape an end-to-end intelligent system
Help maintain a high research bar and push Astrus toward breakthrough capability
WHO YOU ARE
WHO YOU ARE
WHO YOU ARE
You’re deeply aligned with our values and excited about AI for hardware
You’ve built and trained custom neural networks in novel domains — not just used off-the-shelf models
You have experience with large-scale deep learning (e.g. transformers, diffusion, or graph models)
You bring strong research instincts: curiosity, experimental thinking, and rigor
You’re comfortable writing clean well structured research code
You work with engineers to transition research code to production including collaborating to define clean interfaces and understanding requirements
You have a strong foundation in Python and deep learning frameworks such as JAX, PyTorch or TensorFlow; bonus for experience with JAX, Ray, or GCP
MS or PhD in a relevant field (or similar research experience)
You’re deeply aligned with our values and excited about AI for hardware
You’ve built and trained custom neural networks in novel domains — not just used off-the-shelf models
You have experience with large-scale deep learning (e.g. transformers, diffusion, or graph models)
You bring strong research instincts: curiosity, experimental thinking, and rigor
You’re comfortable writing clean well structured research code
You work with engineers to transition research code to production including collaborating to define clean interfaces and understanding requirements
You have a strong foundation in Python and deep learning frameworks such as JAX, PyTorch or TensorFlow; bonus for experience with JAX, Ray, or GCP
MS or PhD in a relevant field (or similar research experience)
WHY THIS ROLE MATTERS
WHY THIS ROLE MATTERS
WHY THIS ROLE MATTERS
You’ll be joining a small, technical, high-talent-density team tackling a problem that deeply matters. Our ambition is enormous, and we’re just getting started.
This role comes with generous equity, full benefits, and the opportunity to lead at the frontier of deep learning and physical design.
You’ll be joining a small, technical, high-talent-density team tackling a problem that deeply matters. Our ambition is enormous, and we’re just getting started.
This role comes with generous equity, full benefits, and the opportunity to lead at the frontier of deep learning and physical design.
Ready to radically improve global computation? 🚀📈🌎 🤖
Ready to radically improve global computation? 🚀📈🌎 🤖
Reach out to careers@astrus.ai or Steph Hector for more details
Reach out to careers@astrus.ai or Steph Hector for more details