1. X
  2. Siddhartha Srinivasa
Log inSign up
Siddhartha Srinivasa
571 posts
Siddhartha Srinivasa profile banner
user avatar

Siddhartha Srinivasa

@siddhss5
Boeing Endowed Professor in #robotics @uwcse. I’m a full-stack roboticist, tennis player, book reader, & cook. Tweets my own.
Seattle, WA
goodrobot.ai
Joined April 2008
16
Following
4,445
Followers
RepliesRepliesMediaMedia
  • user avatar
    Siddhartha Srinivasa
    @siddhss5
    Jun 30
    I had a blast speaking about deploying robots in the real world with @BradPorter_ on his Deployed podcast!
    user avatar
    Brad Porter
    @BradPorter_
    Jun 29
    “We humans are a local minimum in evolution.” @siddhss5 built humanoids for DARPA. His case against them is math: two arms and five fingers is just where evolution landed. Roboticists can build whatever actually works. Deployed EP8 drops tomorrow.
    Image
    00:00
  • user avatar
    Siddhartha Srinivasa
    @siddhss5
    Jun 4
    It’s time to tear this glass house down and rebuild. Are you guys gonna make me come out of retirement for this @Ken_Goldberg?
    user avatar
    Ken Goldberg
    @Ken_Goldberg
    Jun 4
    Correction: two non-surgical robot arms demonstrate a single stitch with zero experimental context after disparaging real research results as “brittle”: stitch-2.github.io
  • user avatar
    Siddhartha Srinivasa
    @siddhss5
    Jun 2
    Post more videos of your robots using ssik! Love it!
    user avatar
    Tom Silver
    @tomssilver
    Jun 2
    Gave ssik a try on our new Vega U from @DexmateAI. Seems to work great!
    Image
    00:00
  • user avatar
    Siddhartha Srinivasa
    @siddhss5
    Jun 1
    Fifteen years ago, I had the privilege of working with Rosen Diankov at CMU on his PhD thesis. The capstone was IKFast — for a generation of roboticists, the definition of what analytical inverse kinematics could be. Today, I'm excited to release the next chapter: ssik. 1/
  • user avatar
    Siddhartha Srinivasa
    @siddhss5
    Jun 19, 2025
    I’ve been recently thinking much about data-efficiency in learning. While there’s no way to overcome the quadratic increase in mistakes with noninteractive IL like BC, a good representation that’s both attentive and generalizable significantly improves the constant.
    user avatar
    Yunchu Zhang
    @yunchuzh
    Jun 19, 2025
    How should a robot perceive the world? What kind of visual representation leads to robust visuomotor policy learning for robotics? Policies trained on raw images are often fragile—easily broken by lighting, clutter, or object variations—making it challenging to deploy policies
    Image
    00:00

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Advertisement
Advertisement