Abhishek Harshvardhan Mishra

writes here as tokenbender

ML researcher focused on efficient training, reinforcement learning research, agentic workflows, and personal systems for doing better work over long horizons.

Portrait of Abhishek Harshvardhan Mishra

this is tokenbender's website.

i write about efficient training and reinforcement learning research, new agentic research workflows, and personal frameworks for doing better work over long horizons.

guide: start with welcome for context, browse everything in archive, or jump directly into a category below.

navigation: use archive by topic when you want themes, and archive by date when you want chronology.

research

2
see all ->

technical

2
see all ->

personal

2
see all ->