/vedantasp

My name is Vedanta, I am
a student (and) researcher.

My name is Vedanta, I am
a student (and) researcher.

My name is Vedanta, I am
a student (and) researcher.

Currently, A research collaborator at Meta, working with discrete diffusion models. I am among the top 3000 students chosen for Amazon’s ML Summer School and also an accepted candidate for Oxford's ML Summer School 2025.

Currently, A research collaborator at Meta, working with discrete diffusion models. I am among the top 3000 students chosen for Amazon’s ML Summer School and also an accepted candidate for Oxford's ML Summer School 2025.

Currently, A research collaborator at Meta, working with discrete diffusion models. I am among the top 3000 students chosen for Amazon’s ML Summer School and also an accepted candidate for Oxford's ML Summer School 2025.

Github

Twitter

Blogs

CV

I always struggle a bit with I'm asked about the "hallucination problem" in LLMs. Because, in some sense, hallucination is all LLMs do. They are dream machines.

I always struggle a bit with I'm asked about the "hallucination problem" in LLMs. Because, in some sense, hallucination is all LLMs do. They are dream machines.

I always struggle a bit with I'm asked about the "hallucination problem" in LLMs. Because, in some sense, hallucination is all LLMs do. They are dream machines.

research experience

research experience

research experience

I initiated my research journey during my initial year as a member of an IoT Research Group, co-authoring 2 papers. Subsequently, I have immersed myself in the study and analysis of research papers on LLMs, with a focus on improving their performance and efficiency. I also served as a research intern at the International Institute of Technology, Bangalore and ended up publishing a paper in UBMK 2024. Currently, a member of Precog - IIIT Hyderabad.

I initiated my research journey during my initial year as a member of an IoT Research Group, co-authoring 2 papers. Subsequently, I have immersed myself in the study and analysis of research papers on LLMs, with a focus on improving their performance and efficiency. I also served as a research intern at the International Institute of Technology, Bangalore and ended up publishing a paper in UBMK 2024. Currently, a member of Precog - IIIT Hyderabad.

Projects

Projects

tinygrad

This project is an enhanced version of the micrograd neural network library, inspired by Andrej Karpathy's micrograd. It provides a simple, educational implementation of backpropagation and neural networks from scratch in Python.

dRAGarys

A self-hosted RAG system leveraging LangChain and LangGraph to enhance document understanding

and question-answering capabilities.
Implemented LangChain’s retrieval-augmented generation framework to improve the accuracy of responses by

combining retrieval and generation steps

petrichor

Developed a mental disorder prediction model using DistilBERT, achieving 97% accuracy on test data.
Synthesised a diverse dataset of 1000+ posts from subReddits related to various psychological disorders and medical

publications using PRAW and GPT-4o.

my board

created a nice-ish looking website using bolt with supabase for my daily tasks,
has nice animations and changes colors acc to time of the day

exploring cursor

created a chatbot with groq as the backend using cursor, also react is fun

aTaskofSimilarities

futile and successful attempts in understanding textual and phrase similarity starting from square one,
attaining higher accuracy than publically available models.

Let's get to know each other.

Let's get to know each other.

Let's get to know each other.

Get in touch.

Get in touch.

Get in touch.