home
Choose the desired section from the header to navigate
home is the default state.
rl, robotics, world models, embodied ai, and system stuff.
im a cs undergrad. im interested in embodied ai (rl as a focus), autonomus robotics, decentralized systems and pretty much any topics related to these. there is only one goal that i want to achieve, work towards making robotic brains autonomous, provide them autonomy, i want to have a humanoid assistant some day and want to work towards that. i love rl a LOT, i see huge huge potential to achieve my desires of an autonomous robotic assistant through this. i see this as my path there. lastly, i love anime, its my lifeblood at this point to keep moving forward. :) i am always open to debate, discussion, and collaboration, regarding any of the topic above, even if they are remotely related, you can be sure i will make my way there :))
i am not sure what is with this image, but i feel its beautiful, and love the image
now
what im on right now
main areas im spending time on right now.
- reinforcement learning learning everything about rl from scratch, all content from 20th century to the 21st century, following Aske Plaat's deep rl graduate textbook as of now.. also spending quite a bit of time experimenting with various other options too :)
- LLMs learning everything about llms, again from back in the time.. including stuff all the way from bag of words to transformers, been hearing that the chinese folks are coming up with something wild, gotta look into that too.
- robotic simulations working towards getting myself familiar with various robotic simulation platforms and tools... purely hands on, no course, no structure.. will try, fk around and find out.. playing around with (mujoco, matlabs simulink/simscape, isaac lab, gazebo, ros, all open robotics products)
- world models trying to play with world models, ive got some ideas i want to try and get working, might be real useful for robots training thorugh rl
stack
skill tree / sub stuff
this section is something ive always struggled with, when exactly could one call that they have acquired a certain skill, im sure the field keeps evolving... so its better to judge me by having a talk. also what if i learned something but forgot it, or learned it but cant apply it.
languages + base stuff
-
core
- python
- embedded c/c++
- java
- javascript
-
web + model formats
- html
- css
- urdf
- mjcf
ai / ml / learning side
-
concepts
- supervised learning
- unsupervised learning
- reinforcement learning
- deep learning
- llm basics, transformers, encoders, decoders, rnns
-
tools
- numpy
- pandas
- matplotlib
- seaborn
- scikit-learn
- gymnasium
- stable baselines
- hugging face
- actually there are a lot more, simply cannot be listed here
robotics / sim / embedded side
-
simulation
- matlab
- simulink
- simscape multibody
- mujoco
- gazebo
- isaac lab
- ros
-
hardware
- arduino
- raspberry pi
- sensor work
- embedded c from robotics training
- i also played by dismantling and playing around with my personal systems, if that is of any significance (all the way to the cpu)
infra / build / distributed side
-
workflow
- git
- github
- github actions
- jupyter
- colab
- bash
-
systems
- docker
- fastapi
- parallax for decentralized heterogeneous inference
- blockchain architecture, dapps, daos, tokenomics (blockchain had its time, was a good time of my life)
maybe more skills, ive decided to not waste anymore time on this section, as its really arbitrary, and skills can be learned and will change with time.
path
where this is going
what does the path look like.
-
right now
cs undergrad, figuring things out, forming connections, experimenting.
-
masters + research + maybe some entrepreneurship
nothing much to explain, start working towards a masters, do research in the areas mentioned above, publish papers, contribute to open source, and keep building stuff, i see a lot of stuff that has to be built... rl has a long way to go, and so does robotics. We are still struggling to replicate stuff a human could do with close to no effort.
-
research + anime
keep working towards that goal of building an autonomous robotic assistant, get it to be as close to a human as possible. playing with emotions seems to be a fun one, play with hormones...
updates
progress updates
pick a note from the list and it opens on the right.
articles
articles
longer writeups.
papers
papers
published research and formal work.
images
gallery
stuff i want to keep around visually.
loading images...
links
links
public links for now.
projects
projects
github repos.
contact
contact me
direct contact points.