Shivakanth

Hello to my small corner of the internet!

Update: Looking for full-time research engineer and research scientist roles in 2026!

I’m a Deep RL Researcher at ARAYA in Tokyo, where I build post-training and deployment systems for vision-language-action models in assistive robotics. My work sits between embodied AI research and systems engineering: I develop reinforcement-learning methods, then profile and optimize the training and inference stack so they can run on real robots.

I completed a Research MSc in Computer Science at Mila, supervised by Prof. Samira Ebrahimi Kahou, with a thesis on sample-efficient reinforcement learning for real-world systems. I was also a research intern with John Langford at Microsoft Research NYC, working on reachability-aware representations for efficient planning.

My research spans offline and online RL, embodied foundation models, representation learning, and human-robot interaction. I completed my undergraduate degree in Instrumentation and Control Engineering at NIT Trichy. Feel free to reach out if you’d like to talk about reinforcement learning, robotics, or making research systems fast enough to be useful.

You can read my CV online or download the PDF.