Rohit Mujumdar

Rohit Mujumdar

Machine Learning Engineer & Researcher

Specialized in AI/ML systems, misinformation detection, and scalable data solutions

Email Resume LinkedIn GitHub Google Scholar

Machine learning engineer and researcher with expertise in building production ML systems and conducting impactful research. MS in Computer Science from Georgia Tech. Experience at Intel, IBM Research, and NCR Corporation developing innovative AI solutions for real-world challenges.

Experience

Intel Corporation

2022 - 2024

Machine Learning Engineer

  • Developed RAG-based GenAI solutions for internal document interaction
  • Applied data science expertise to deliver insights across multiple teams
  • Led end-to-end project implementation from conception to deployment

IBM Research

2019 - 2020

Research Intern

  • Developed predictive models for service request resolution time estimation
  • Contributed to 2 patents on ticket-agent matching and demand forecasting
  • Received Best Poster Award (Honorable Mention) at INFORMS 2020

NCR Corporation

2021 - 2022

Software Engineer

  • University Recruiting Champion
  • Vice President of Public Relations, iNCRedible Toastmasters

Education

Georgia Institute of Technology

Master of Science in Computer Science

Specialization in Machine Learning

Head Teaching Assistant - AI, Ethics and Society (Dr. Ayanna Howard)

Selected Publications

Language Disparity

Overcoming Language Disparity in Online Content Classification with Multimodal Learning

Gaurav Verma, Rohit Mujumdar, Zijie J. Wang, Munmun De Choudhury, Srijan Kumar

ICWSM 2022

Investigated disparity between English and non-English language models, demonstrating how multimodal machine learning incorporating image information can address systematic performance gaps in pre-trained large language models.

Website Paper
HawkEye

HawkEye: A Robust Reputation System for Community-based Counter-Misinformation

Rohit Mujumdar, Srijan Kumar

ASONAM 2021

First research study on Twitter's Community Notes (Birdwatch). Identified vulnerabilities to manipulation attacks and developed HawkEye, a cold-start-aware graph-based algorithm demonstrating improved robustness. Open-sourced implementation influenced improvements to Twitter's algorithm.

Paper Code Video
Ontology

Recognizing Similar Relationships Within Ontology to Fine Tune Ontology

Neelam Chandolikar, Rishav Raj, Rohit Mujumdar

ICDMAI

Proposed solution for identifying similar relationships in automated ontology learning to enable fine-tuning and reduce duplicate relationships.

Paper
INFORMS

A Heuristic Approach To Compute Service Request Resolution Time

Rohit Mujumdar, Pawan Chowdhary, Shubhi Asthana

INFORMS 2020 - Best Poster Award (Honorable Mention)

Statistical analysis and regression-based techniques for predicting incident ticket resolution times, employing dynamic rolling windows and auto-regressive techniques.

Poster Video

Patents

Patent 1

Ticket-agent matching and agent skillset development

Rohit Mujumdar, Shubhi Asthana, Pawan Chowdhary, Aly Megahed, Bing Zhang

US20220270019A1

Patent 2

Demand forecasting of service requests volume

Bing Zhang, Shubhi Asthana, Pawan Chowdhary, Aly Megahed, Rohit Mujumdar, Taiga Nakamura

US20220164744A1

Selected Projects

Fairness

Exploring Fairness in Graph Embeddings

Rohit Mujumdar, Sanjana Garg, Rohit Gajawada

Demonstrated bias in graph embeddings for movie recommendation systems and introduced fairness mitigation methods based on Fairwalk.

Report Code
Content Moderation

Explainable Content Moderation Using CNNs

Rohit Mujumdar, Shalini Chaudhuri, Sreehari Sreejith, Sushmita Singh

Built content moderating system to identify and flag violent/gory content achieving 0.89 accuracy using Transfer Learning with VGG-16. Implemented Grad-CAM for model explainability.

Report Code Video
MAG

Do Scientific Ideas Originating from Prestigious Universities Spread Faster?

Rohit Mujumdar, David Kartchner

Investigated imbalance in spread of ideas across academic research networks using disease spread models adapted from epidemiology.

Website Report Code