Dr. Tanmay Sen

Dr. Tanmay Sen

Assistant Professor
SQC & OR Unit, Indian Statistical Institute Kolkata
Room No- 613, 6th floor, Kolmogorov Bhavan, 203 B. T. Road,
Kolkata - 700108, India
Email: sentanmay518@gmail.com / tanmay.sen@isical.ac.in
Phone: +91 8084526317 / +91 6363168171

About

I am an Assistant Professor in the SQC & OR Unit at the Indian Statistical Institute, Kolkata. My research interests lie at the intersection of Survival Analysis, Reliability, Deep Learning, Machine Learning, Time Series Anomaly Detection, Log Analysis, NLP, LLM.

Prior to joining ISI, I worked as a Lead Data Scientist at Ericsson India Global Services, where I led several machine learning and AI initiatives focused on telecommunications networks, anomaly detection, and federated learning systems. I have extensive experience in developing scalable ML solutions for industrial applications.

I completed my PhD in Mathematics from IIT Patna in 2019, where I worked on statistical inference and optimal life testing plans under various censoring schemes. My current research focuses on federated learning, multi-task NLP, multi-modal LLM, anomaly detection, log analysis and survival analysis for big data applications.

Flag Counter

Education

  • PhD in Mathematics, IIT Patna, India (2019)
    Thesis: Statistical Inference and Optimal Plans Under Hybrid Censoring Scheme
    Advisor: Dr. Yogesh Mani Tripathi
  • M.Tech in Mathematics and Computing, IIT Patna, India (2014)
    Thesis: Multi-Objective Portfolio Optimization Using Evolutionary Algorithms
    Advisors: Dr. Sriparna Saha (IIT Patna), Prof. Arnab Kumar Laha (IIM Ahmedabad)
  • M.Sc. in Pure Mathematics, University of Calcutta, India (2011)
  • B.Sc. in Mathematics (Hons.), University of Calcutta, India (2009)

Experience

Ericsson (Kolkata, India)

Lead Data Scientist (Aug 2019 – Oct 2024)

  • Unsupervised KPI anomaly detection using ML/DL models (Prophet, seq-seq), Pyspark, Docker.
  • Ping-pong handover anomaly detection & prediction.
  • Code translation (Java to Go) using OpenAI GPT-4.
  • Root cause analysis of CI logs using drain3, deployed with Docker.
  • LLM-based test case generation with RAG; used llama3, MIXTRAL, reranker, various embeddings, PDF/CSV parsers.

Data Scientist

  • Federated Learning: Class imbalance-aware aggregation for KPI degradation classification.
  • Power outage prediction using DL on structured outage data.
  • Log analysis using ELK stack, heuristic parsers, and BERT-based models.
  • Churn classification using H2O, MLflow, SHAP, LIME.

L&T Technology Services (Bangalore, India)

Senior Engineer (Jun 2018 – Jul 2019)

  • Predictive maintenance for CNC tools, time series (LMP) forecasting via LSTM/ARIMA.
  • Vehicle routing optimization using GA.
  • Log anomaly detection using autoencoders and ML; authored white paper.
  • Developed analytics module for LTTS AIKNO framework.

Yatra Online Pvt Ltd (Gurgaon, India)

Intern (Research) (Jan 2018 – Apr 2018)

  • Built chatbot using LSTM for enhanced human interaction on web/mobile app.

Publications

Journal Articles

Federated Multitask Learning for Complaint Identification Using Graph Attention Network
A Singh, S Chandrasekar, T Sen, S Saha
IEEE Transactions on Artificial Intelligence, 2023
Mtbullygnn: a graph neural network-based multitask framework for cyberbullying detection
K Maity, T Sen, S Saha, P Bhattacharyya
IEEE Transactions on Computational Social Systems, 2022
Statistical inference and Bayesian optimal life‐testing plans under Type‐II unified hybrid censoring scheme
T Sen, R Bhattacharya, B Pradhan, YM Tripathi
Quality and Reliability Engineering International, 2021
Inference and optimum life testing plans based on Type-II progressive hybrid censored generalized exponential data
T Sen, R Bhattacharya, YM Tripathi, B Pradhan
Communications in Statistics-Simulation and Computation, 2020
On progressively censored inverted exponentiated Rayleigh distribution
RK Maurya, YM Tripathi, T Sen, MK Rastogi
Journal of Statistical Computation and Simulation, 2019
Inference for an inverted exponentiated Pareto distribution under progressive censoring
RK Maurya, YM Tripathi, T Sen, MK Rastogi
Journal of Statistical Theory and Practice, 2019
Inference for kumaraswamy distribution based on type I progressive hybrid censoring
F Sultana, YM Tripathi, SJ Wu, T Sen
Annals of Data Science, 2022
Statistical inference and optimum life testing plans under Type-II hybrid censoring scheme
T Sen, YM Tripathi, R Bhattacharya
Annals of Data Science, 2018
Statistical inference for Lognormal distribution with type-I progressive hybrid censored data
T Sen, S Singh, YM Tripathi
American Journal of Mathematical and Management Sciences, 2019
Generalized hybrid censored reliability acceptance sampling plans for the Weibull distribution
T Sen, R Bhattacharya, YM Tripathi, B Pradhan
American Journal of Mathematical and Management Sciences, 2018
Determination of Bayesian optimal warranty length under Type-II unified hybrid censoring scheme
T Sen, R Bhattacharya, B Pradhan, YM Tripathi
Quality Technology & Quantitative Management, 2022
Fisher information in generalized progressive hybrid-censored data
T Sen, B Pradhan, YM Tripathi, R Bhattacharya
Statistics, 2018
Quantile estimation for a progressively censored exponential distribution
YM Tripathi, C Petropoulos, T Sen
Communications in Statistics-Theory and Methods, 2020

Conference Papers

Grafehty: Graph neural network using federated learning for human activity recognition
A Sarkar, T Sen, AK Roy
IEEE ICMLA, 2021
Federated multi-task learning for complaint identification from social media data
A Singh, T Sen, S Saha, M Hasanuzzaman
ACM Hypertext and Social Media, 2021
Bi-objective portfolio optimization using archive multi-objective simulated annealing
T Sen, S Saha, A Ekbal, AK Laha
HPC Applications Conference, 2014
Federated Meta-Learning for Emotion and Sentiment Aware Multi-modal Complaint Identification
A Singh, S Chandrasekar, S Saha, T Sen
EMNLP, 2023
Multivariate Missing Value Imputation using Weighted Cluster Softmax Technique
A Maji, S Amed, T Sen
Interdisciplinary Research in Technology and Management, 2021

Preprints & arXiv

MHARFedLLM: Multimodal Human Activity Recognition Using Federated Large Language Model
A Bandyopadhyay, R Basu, T Sen, S Das
arXiv, 2025
LogTinyLLM: Tiny Large Language Models Based Contextual Log Anomaly Detection
IT Ocansey, R Bhattacharya, T Sen
arXiv, 2025
Determination of Optimum Warranty Region for Two Dimensional Dependent Data
R Das, T Sen, R Bhattacharya, B Pradhan
arXiv, 2025
Hatetinyllm: hate speech detection using tiny large language models
T Sen, A Das, M Sen
arXiv, 2024
A Greedy Approach for Offering to Telecom Subscribers
PK Bhunre, T Sen, A Sarkar
arXiv, 2023
LogAnMeta: Log Anomaly Detection Using Meta Learning
A Sarkar, T Sen, S Kundu, A Sarkar, A Wazed
arXiv, 2022

Patents

INTELLIGENT SOURCE SWITCHING IN MICROGRID CONNECTED HOUSEHOLDS
A Sarkar, A Sen, AK Roy, DN Hegde, H Hallberg, T Sen
US Patent App. 18/688,403, 2024

Professional Activities

I am serving as a reviewer of the following journals:

  • Neural Networks
  • PLOS ONE
  • Expert Systems with Applications
  • IEEE Internet Computing
  • IEEE Transactions on Computational Social Systems
  • Communications in Statistics-Simulation and Computation

Research Interests

  • Reliability and Survival Analysis
  • LLM & GenAI
  • Federated Learning
  • Graph Neural Network
  • Meta Learning
  • Anomaly Detection and Prediction
  • Log Analysis
  • Time Series
  • Multi-task NLP
  • Multi-objective optimization (Portfolio / Vehicle routing etc) using meta-heuristic algorithms

Collaborators

  • Prof. Sriparna Saha (IIT Patna)
  • Prof. Arnab Kumar Laha (IIM Ahmedabad)
  • Dr. Ritwik Bhattacharya, Department of Mathematical Sciences, University of Texas at El Paso
  • Prof. Biswabrata Pradhan, SQCOR unit, ISI Kolkata
  • Dr. Yogesh Mani Tripathi, Department of Mathematics, IIT Patna

Students

M.Tech Thesis Guided

  • Abhilash Kar (2025) - Time Series Anomaly Detection using Transformer Models, M.Tech QROR, ISI Kolkata
  • Arunava Phadikar (2025) - LLM For Hate Speech Detection, M.Tech QROR, ISI Kolkata
  • Sounak Biswas (2025) - Survival Analysis using Deep Learning, M.Tech QROR, ISI Kolkata
  • Deepak Pal (2025) - Multimodal Chatbot for Skin Disease Diagnosis, M.Tech QROR, ISI Kolkata

Bachelor Thesis Guided

  • Asmit Bandyopadhyay (2025) - Multimodal Human Activity Recognition Using Federated Large Language Model (Dept. of Computer Science, IEM, Kolkata)
  • Dnyaneshwari Ghare (2025) - Medical Supply Chain Optimization, Dept. of Industrial Engineering, IIT, Kharagpur

Teaching

  • Reliability 1, M.Tech QROR 2nd Semester (Completed)
  • Reliability 2 (Jointly with Prof. Biswabrata Pradhan), M.Tech QROR 3rd Semester (Ongoing)
  • Business Analytics (Jointly with Prof. Prasun Das), M.Tech QROR 3rd Semester (Ongoing)