Skip to content
View rockysaikia730's full-sized avatar
πŸ‘‹
πŸ‘‹

Highlights

  • Pro

Block or report rockysaikia730

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
rockysaikia730/README.md

πŸ‘‹ Hi there, I am A. Rockey Saikia (he/him)

  • πŸŽ“ MSc student in Computational Science and Engineering at EPFL, Lausanne
  • πŸ‘¨β€πŸ’» DAAD-WISE Fellow at Technical University of Munich
  • πŸŽ“ B.Tech in Engineering Physics at Indian Institute of Technology, Roorkee
  • πŸ’» Machine Learning, Maths, Software Development are my interests
  • 🚞 I like hiking, football and food!

πŸ“« Let's connect!

CV Mail LinkedIn GitHub Google Scholar

πŸ”¨ Few of my projects!

  • European Option Pricing: A Monte Carlo estimator to predict payoffs of European options, including both path-dependent and path-independent options. We implement crude monte carlo, adaptive importance sampling and importance sampling via optimal control to reduce variance to simulate rare events and accurate pricing.
  • Neural Audio Codec: Implemented residual vector quantization using 1D CNNs based on the Soundstream architecture to compress signals into discrete latent embeddings for high-fidelity reconstruction.
  • Gene Guard: Gene Guard is a Python tool designed to parse and flag data files containing genomic data using machine learning (NLP, CNN) and rule-based algorithms. Beyond simple detection, it is capable of identifying and extracting genetic sequences embedded within mixed or obfuscated text to prevent genomic data leakage.
  • RAG: Developed a multimodal Retrieval-Augmented Generation (RAG) application with LlamaIndex, Docling and Streamlit for prcise retrievals using sentence level chunking, metadata, query augmentation.
  • Benchmarking Optimizers: Benchmarked Adam, Muon, Scion and Dion optimizers to train Vision Transformers and ResNet on CIFAR-10 dataset implemented from scratch using PyTorch.
  • ODE Solver: A C++ based ODE Solver for dynamic rank tensors using both implicit and explicit ODE solving methods (Runge Kutta and Linear Multistep methods).
  • MyPi: MyPi is an over-engineered plug and play device which can convert any USB based Mouse(Wired or Wireless) to an IP Mouse. An IP Mouse is basically an Interent Protocol device which can connect to your PC over a WiFi connection and can control your device wirelessly in real time. It uses multithreading and socket programming.

Pinned Loading

  1. AudioTokenizer AudioTokenizer Public

    Jupyter Notebook

  2. GeneGuard GeneGuard Public

    Jupyter Notebook 1

  3. Optimizers Optimizers Public

    Jupyter Notebook

  4. RAG RAG Public

    Python

  5. MyPi MyPi Public

    Python 2