Adithya "Adi" Yerramsetty

I'm currently

• a junior at ASU going for a double major in CS + Math
• working on building a NeRF of all of Seattle from some driving data I collected
• implementing and playing with various algos/problems in ML/CV including Diffusion Models, NeRFs, Visual-SLAM and more

In the past I

• interned as a Deep Learning Intern at Silicon Valley Bank(May 2022 - January 2023), where I worked on Graph Neural Nets for investment predictions
• was president at the ASU Machine Learning Club, where I taught ML, Deep Learning and more to interested members
• was an undergrad researcher at ASU DREAM Lab working on NeRF/View Synthesis
• won at some hackathons, including DataFest @ ASU, ASU AI In Education and SVB @ ASU 2021

For fun, I've been working on

• a toy scalar-only AutoGrad library I built, inspired by MicroGrad by Karpathy - not very performant, but can fit Linear Models and MNIST MLP, and in theory can handle most models
• SimpleVAE: some simple VAEs I've trained up, as stepping stones towards building an LDM for NeRF supervision.
• Linear Regression in OCAML: out of curiosity about how OCAML works, I just built a Linear Regression model in it. As it turns out, OCAML is pretty nice.

I'm currently writing up

• a whirlwind tour of ML, going over some of the topics we would normally cover at the ML club
• an explainer for the proof for gradient descent; a lot of it is detailed nicely by this article, but I want to go through it myself
• a derivation for the SVD Solution to Homogenous LLS. I have it on my old site, but haven't moved it here yet. It's pretty useful in V-SLAM, from triangulation to Fundamental/Essential Matrix computations
• a post on the basics of vector search(LSH, HNSW, etc). They're a super cool application of Neural Nets, and fascinated me during my freshman year