Deepon HalderML RESEARCH / SYSTEMS ENGINEERING

WORK

Projects & publications

01

PaperCode

A practice platform where people implement machine-learning papers, validate ideas with tests, track progress, and use a job stack built for ML roles.

Year
2026 – Present
Stack
Next.js / ML Platform
Role
Co-creator & Engineer
  • 20+ paid users
  • 200+ daily submissions
Launch platform
PaperCode machine-learning practice platform
02

Solving ML Papers

A growing collection of from-scratch implementations that turns deep-learning papers and core ideas, from attention and BERT to diffusion and decoding, into working code.

Year
2024 – Present
Stack
Python / Jupyter
Role
Creator & Maintainer
  • 19 implementations
  • Research-to-code
Explore repository
Solving ML Papers notebooks and implementation work
03

Neugrad

A lightweight autograd engine with tensor operations, automatic differentiation, neural-network layers, optimizers, and backpropagation, built to understand PyTorch from first principles.

Year
2024
Stack
Python / NumPy
Role
Creator & Developer
  • Autograd from scratch
  • MIT licensed
Explore repository
Neugrad autograd engine and computational graph
[08 Papers]

PUBLICATIONS

Language, reasoning, evaluation, and learning systems.

  1. 01
    2026INTERSPEECH 2026: Main Conference

    Preferences of a Voice-First Nation: Large-Scale Pairwise Evaluation and Preference Analysis for TTS in Indian Languages

    Srija Anand, Ashwin Sankar, Ishvinder Sethi, Aaditya Pareek, Kartik Rajput, Gaurav Yadav, Nikhil Narasimhan, Adish Pandya, Deepon Halder, Mohammed Safi Ur Rahman Khan, Praveen S V, Shobhit Banga, Mitesh M Khapra

    [arXiv] ↗
  2. 02
    2026Preprint

    Top-b: Entropic Regulation of Relative Probability Bands in Autoregressive Language Processes

    Deepon Halder, Raj Dabre

    [arXiv] ↗
  3. 03
    2026Preprint

    Multilingual TinyStories: A Synthetic Combinatorial Corpus of Indic Children’s Stories for Training Small Language Models

    Deepon Halder, Angira Mukherjee

    [arXiv] ↗
  4. 04
    2026ICBINB Workshop @ ICLR 2026

    Not All Time Is Gregorian: Evaluating LLMs on Cultural Calendar Systems

    Deepon Halder, Adish Pandya, Raj Dabre

  5. 05
    2026ACL Findings 2026

    Scripts Through Time: A Survey of the Evolving Role of Transliteration in NLP

    Thanmay Jayakumar, Deepon Halder, Raj Dabre

    [Project] ↗
  6. 06
    2025EACL Findings 2026

    RiddleBench: A New Generative Reasoning Benchmark for LLMs

    Deepon Halder, Alan Saji, Thanmay Jayakumar, Raj Dabre, Ratish Puduppully, Anoop Kunchukuttan

    [arXiv] ↗[Project] ↗[Dataset] ↗
  7. 07
    2025MELT Workshop @ COLM / WAT @ IJCNLP-AACL 2025

    CycleDistill: Bootstrapping Machine Translation using LLMs with Cyclical Distillation

    Deepon Halder, Thanmay Jayakumar, Raj Dabre

    [arXiv] ↗[Project] ↗[Code] ↗
  8. 08
    2025Survey Paper

    A Comprehensive Survey of Data Poisoning Attacks

    Deepon Halder, Anshika Gupta, Diya Ghosh, Hafizur Rehman

    [Paper] ↗