Rahul Goswami

Rahul Goswami

Ph.D. Mathematics, IIT Guwahati

Supervised by Dr. Palash Ghosh (IIT Guwahati) and Dr. Tanujit Chakraborty (Sorbonne University Abu Dhabi)

I am a doctoral researcher in the Department of Mathematics at the Indian Institute of Technology Guwahati, and an affiliate of SAFIR at Sorbonne University Abu Dhabi. My work sits between statistics and machine learning: I develop methods for survival analysis, tree-based ensembles, and time-series forecasting, and release them as open-source software.

Recent work includes boosted trees for time-varying relative risk (NeurIPS 2026), discrete-time neural survival models with guaranteed monotonicity (TMLR), kernel ensembles for patient-specific survival prediction (ACML 2026), and long-memory reservoir computing for data-scarce dengue forecasting.

Research interests: survival analysis · ensemble methods · probabilistic and long-memory forecasting · interpretable machine learning

Read my research blog at statml.in

UpdatesNews

Oct 2026
LAIR-Net preprint on arXiv, with a project site at lairnet.statml.in.
2026
CCTimeBoost accepted at NeurIPS 2026.
2026
KESurv accepted at ACML 2026 (website).
2026
Deep Discrete-Time Survival Analysis with Guaranteed Monotonicity accepted at TMLR.
Jul 2026
Dengue forecasting preprint on arXiv; companion library memory-esn released.
Mar 2026
EnTransformer preprint on arXiv.

ResearchPublications

  1. AcceptedCCTimeBoost: Learning Time-Varying Relative Risk with Case-Control Boosted Trees.
    R. Goswami (sole author). NeurIPS 2026, in press.
  2. AcceptedDeep Discrete-Time Survival Analysis with Guaranteed Monotonicity.
    R. Goswami (sole author). Transactions on Machine Learning Research (TMLR), in press. Formerly circulated as KMNet.
    Code
  3. AcceptedKESurv: A Kernel Ensemble Method for Patient-Specific Survival Prediction.
    R. Goswami (sole author). ACML 2026, in press.
    Website
  4. PublishedMART: Moving Average Randomized Tree.
    V. K. Giri and R. Goswami (equal first authors). Machine Learning (Springer).
    Paper · Code
  5. PublishedShape Penalized Decision Forests for Imbalanced Data Classification.
    IEEE.
    Paper · PyPI
  6. Under reviewLong-Memory Reservoir Computing for Data-Scarce Dengue Forecasting.
    R. Goswami, S. Paul, P. Ghosh, T. Chakraborty. TMLR, under review.
    arXiv:2607.11272 · Website
  7. Under reviewEnTransformer: A Deep Generative Transformer for Multivariate Probabilistic Forecasting.
    R. Goswami (first co-author), R. Pathak, M. Panja, P. Ghosh, T. Chakraborty. New England Journal of Statistics in Data Science, under review.
    arXiv:2603.07108 · Code · Docs
  8. PreprintLAIR-Net: Leaky Alignment-Impulse Residual Networks for Tabular Regression.
    R. Goswami, A. Bhambu, B. Karmakar. 2026.
    arXiv:2610.11538 · Website · Code
  9. PreprintConcordance-based Survival COBRA with Regression-Type Weak Learners.
    arXiv:2209.11919
  10. PreprintIntegrated Brier Score based Survival COBRA — A Regression-Based Approach.
    arXiv:2210.12006

Open sourceSoftware

lairnet — Leaky Alignment-Impulse Residual Networks2026Fast leaky alignment-impulse randomized networks for tabular learning; companion to the LAIR-Net paper.PyPI · Website · GitHub
memory-esn — Long-Memory Echo State Networks2026Reservoir-computing library (fESN, wESN) with Cython acceleration; companion to the dengue forecasting paper.PyPI · Website · GitHub
MART — Moving Average Randomized Trees2026Reference implementation of the Machine Learning paper; EMA spans embedded in the split criterion.GitHub
imbalanced-spdf — Shape Penalized Decision Forests2025Imbalanced classification (SPBoDF / SPBaDF) with surface-to-volume regularization.PyPI · Docs
fastkme — Fast Kaplan–Meier Estimator2024Accelerated Kaplan–Meier estimation with weighted curves.GitHub · Blog post

Industry & collaborationExperience

Adobe, NoidaMay – Aug 2024Ph.D. Media & Data Science Research InternStudied memorization, reasoning and counting in large language models; designed quantitative metrics that guided redesign of in-house LLMs.
Huntington's Disease Project (with University of Iowa collaborators)Dec 2022 – presentResearch CollaboratorLinear mixed-effects models for covariates affecting years-to-onset; led data analysis and interpretation.
ByteDanceMar – May 2025Math Expert — reviewer of Ph.D.-level mathematics questions
TuringAug 2024 – Feb 2025AI Trainer & Ph.D. Math Expert — training data and evaluations for advanced AI systems
Allen Institute for AI (AI2)Mar – May 2024Freelance NLP Expert — NER, relation extraction and summarization over historical research corpora
Ready TensorMar – Dec 2021Machine Learning Engineer — production-ready forecasting and classification model templates
Chegg Inc.Jun 2018 – Feb 2021Subject Matter Expert, Statistics & Probability — assisted 1000+ students

ServiceTeaching

  • Head Teaching Assistant (lead for four TAs), Coursera AI Program DA112
  • Teaching Assistant, Financial Risk Analysis, IIT Guwahati (with Dr. Arabin Kumar Dey)
  • Teaching Assistant, Advanced Statistical Algorithms, IIT Guwahati (with Dr. Arabin Kumar Dey)
  • Teaching Assistant, R Programming (UG), IIT Guwahati
  • Teaching Assistant, Applied Statistics (with Dr. Palash Ghosh)

TrainingEducation

Ph.D. MathematicsIndian Institute of Technology Guwahati2021–2026
MITx MicroMasters in Finance (in progress)Massachusetts Institute of Technology (online)2024–present
M.Sc. Statistics and ComputingBanaras Hindu University2018–2020
B.Sc. Statistics & Computer ScienceDDU Gorakhpur University2015–2018

RecognitionHonours

  • GATE (Statistics) — All India Rank 63
  • IIT-JAM (Mathematical Sciences) — All India Rank 306
  • NPTEL Operations Research — Top 1%
  • Institute of Actuaries of India — CS1

ToolsSkills

Programming
Python, R, Cython, Julia, Lean, C
ML frameworks
PyTorch, TensorFlow, JAX, scikit-learn, scikit-survival, Hugging Face Transformers
Software practice
PyPI packaging, Sphinx / ReadTheDocs, pytest, Docker, Git