Rabindra Nath Nandi

Rabindra Nath Nandi

12+ years in Machine Learning, NLP, and AI product engineering

Principal Machine Learning Engineer

ReliSource, Boston, MA, USA

Currently leading ML initiatives at ReliSource, focusing on scalable, production-grade AI solutions for global clients.
Previously, Lead NLP Engineer at Hishab Ltd, building LLM-based customer care and voicebot systems for millions of users.
Passionate about low-latency, robust AI for real-world impact, especially in low-resource languages.
BSc in Computer Science & Engineering, KUET (2014)

Research Interests

Large Language Models (LLMs), Multilingual NLP, Speech Processing
Low-resource Language Technologies
Conversational AI, Dialogue Systems
Information Retrieval, Text Mining
AI for Social Good

Publications

Shahriar Kabir Nahin, RN Nandi, Sagor Sarker, Firoj Alam et al.
TituLLMs: A Family of Bangla LLMs with Comprehensive Benchmarking. Feb, 2025
ACL 2025
Mohammad Jahid Ibna Basher Md Kowsher, Md Saiful Islam, Rabindra Nath Nandi et al.
BnTTS: Few-Shot Speaker Adaptation in Low-Resource Setting
NAACL 2025
Rabindra Nath Nandi, Suman Maity, Brian Uzzi, Sourav Medya.
An Experimental Analysis on Evaluating Patent Citations
EMNLP 2024
RN Nandi, SA Chowdhury, F. Alam.
Pseudo-Labeling for Domain-Agnostic Bangla Automatic Speech Recognition
BLP Workshop@EMNLP2023
RN Nandi, F. Alam.
TeamX at CheckThat! 2023: Multilingual and Multimodal Approach for Check-Worthiness Detection
CLEF 2023
M. Hasanain, A. El-Shangiti, RN Nandi, P. Nakov, F. Alam.
QCRI at SemEval-2023 Task 3: News Genre, Framing, and Persuasion Techniques Detection Using Multilingual Models
SemEval-2023
A. Barrón-Cedeño, F. Alam, T. Caselli, G. Da San Martino, RN Nandi et al.
The clef-2023 checkthat! lab: Checkworthiness, subjectivity, political bias, factuality, and authority
ECIR 2023
RN Nandi, F. Alam, P. Nakov.
Detecting the Role of an Entity in Harmful Memes: Techniques and their Limitations
CONSTRAINT Workshop@ACL2022

See full list of publications →

Updates (more)

11.2025 Invited Speaker in Polyglot Workshop: Developing LLMs for Low-Resource Languages. Polyglot Workshop, University of Bonn
02.2025 Recent Paper on Bangla LLM Pretraining accepted by ACL 2025. Paper Link
01/2025 One paper accepted by NAACL 2025. Paper Link
09/2024 One paper accepted by EMNLP 2024. Paper Link
11/2023 One paper accepted by BLP Workshop@EMNLP2023. Paper Link
03/2023 One paper accepted by SemEval-2023. Paper Link
06/2023 One paper accepted by CLEF 2023 - CheckThat! Lab. Paper Link
12/2022 One paper accepted by ECIR2023. Paper Link
03/2022 One paper accepted by CONSTRAINT Workshop@ACL2022. Paper Link
03/2022 One paper accepted by DravidianLangTech@ACL2022. Paper Link

Projects

Speech and NLP

Rasa based and LLM based voice bot solutions for microfinance organizations, 2023-2024
Features: use-case analysis, prompting, RAG, user data management, reduce latency and hallucination, deployment, real-time monitoring
Tokenizer Build for In-House LLM, 2023
Features: Data collection, deduplication checking, tokenizer built like as gpt-3 and llama tokenizer and check performance.
Bangla Spell Checker, 2023
Features: Data collection, Model Training and Validation
Develop a Text-to-Speech Engine for Bangla Language, 2023
Features: Evaluate state-of-art dataset and models to figure out a product grade TTS System
Slang Detector, 2017
Features: Dataset Collection and Labelling, Model training and validation, API (Java Based)
News Recommendation System, 2016
Features:News Downloader,SubCategory Extraction, Topic Modelling

Computer Vision

EdgeAI based personalized assistant in mobile app, 2021-22
Features: Model inference in on-device for both image, text and heterogenous data, lightweight ML models
FaceAI SDK Development, 2020-2021
Features: Low latency, high accuracy, near realtime inference
Facial Action Unit Detection, 2019
Features: Data Collection tools, Data Labeling, Model training and validation, Deployment for Facial Action Rendering
Parking Area Detection and Illegal Parking Identification from real time CCTV videos, 2019
Features: Data annotation, Object Detection and Re-Identification, Semantic Segmentation, Geometric Algorithm, Video Analysis

Others

Mentoring 20+ engineers and contributed to several projects both independently and as a team member.

Other Activities

Program Committee, Second Workshop on Bangla Language Processing, AACL-IJCNLP 2025.
Program Committee, First Workshop on Bangla Language Processing, EMNLP 2023.
Task Organizer, CheckThat! Lab at CLEF 2023, Greece, 2023.
Reviewer of BLP Workshop @ EMNLP 2023, CLEF2023 - CheckThat! Lab, DravidianLangTech-Workshop@ACL2022