Skip to content
View Rohith-2's full-sized avatar
🧠
BrainStorming
🧠
BrainStorming

Highlights

  • Pro

Block or report Rohith-2

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.

Content in all repositories owned by your account will be closed.
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
Rohith-2/README.md

Rohith Ramakrishnan

LinkedIn Badge GitHub Badge Medium Badge

I build search, retrieval, and NLP systems — some at work, some just because I wanted to know how they worked. Currently a Master's student in Data Science at the University of Michigan; before that, a few years shipping GenAI systems in an aviation engineering org. This page is really about what's in the repos below.

Search & Retrieval

  • bm25-fusion — An ultra-fast BM25 implementation from scratch in Python, JIT-compiled with Numba. Supports classic BM25, BM25+, BM25L, and ATIRE variants, metadata filtering, real-time add/delete, and HDF5 persistence. It's on PyPI — pip install bm25-fusion.
  • url_classification_dl — Feature extraction and engineering on raw URLs, feeding a neural classifier. My most-forked repo, so apparently it's been useful to more people than just me.

Science-flavored side quests

  • Chaos-Game-Representation_BioSeq — Turns DNA sequences into chaos-game fractal images, then compares those images to measure similarity between genes and species. Has a Streamlit GUI and a Medium writeup if you want the full story.
  • SpatioTemporalKoopman — Playing with Koopman-operator / dynamic mode decomposition methods (HODMD) for spatiotemporal systems.
  • big_data_analysis — 20+ years of US flight performance data (1987–2008), crunched in Scala/Spark. A big-data course needed a genuinely big dataset.

For fun

  • ACoin — A toy blockchain/cryptocurrency in Java, built to actually understand block-chaining and proof-of-work rather than just read about them.

Open source contributions

  • LlamaIndex — Triton streaming/chat inference support, and S-BERT re-ranker integration for Hugging Face's Text-Embeddings-Inference
  • Hugging Face
  • NVIDIA Triton Inference Server

GitHub also seems to think this adds up to a Pull Shark and Arctic Code Vault Contributor badge, for what it's worth.

A few papers, if you're into that

  • Analysis of Text-Semantics via Efficient Word Embedding Using Variational Mode Decomposition — PACLIC-35, 2021
  • CT Image Enhancement Using Variational Mode Decomposition for AI-Enabled COVID Classification — 2023
  • Data-Driven Volatile Cryptocurrency Price Forecasting via Variational Mode Decomposition and BiLSTM — ICICC 2022

Connect

Pinned Loading

  1. bm25-fusion bm25-fusion Public

    An ultra-fast BM25 retriever with support for multiple variants and meta-data filtering.

    Python 2 2

  2. url_classification_dl url_classification_dl Public

    URL Feature extraction and Engineering aided with Classification via Neural Networks

    Python 11 11

  3. big_data_analysis big_data_analysis Public

    Performance of Aircraft in the US from 1987 to 2008.

    Scala 1

  4. Chaos-Game-Representation_BioSeq Chaos-Game-Representation_BioSeq Public

    Representation of Bio Sequences via Chaos Game and using the same to find similarities

    Python 8 2

  5. SpatioTemporalKoopman SpatioTemporalKoopman Public

    Forked from ModelsFLOW/HODMD

    HODMD algorithm from Le Clainche & Vega, SIAM J. on Appl. Dyn. Sys. 16(2), 882-925, 2017

    MATLAB 1

  6. ACoin ACoin Public

    A simple implementation of crypto-currency via BlockChain

    Java