Sitemap
A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.
Pages
Posts
experience
Sr ML Engineer
GenAI Platform Team, Quantiphi (Client: AT&T)
Duration: 02 February 2025 - Present
Replacing closed-source LLMs with fine-tuned open-source models in production (saving $800K/year), and building the RL post-training and evaluation infrastructure used by 10+ teams at AT&T.
AI Engineer
AI Engineering Team, BYO
Duration: 08 August 2024 - 02 February 2025
Reproduced TRACE continual learning results on Llama-3.1-8B, and built RAG agents with word/topic guardrails that improved response accuracy by 15%.
AI Consultant
Division of Occupational and Environmental Medicine, Department of Family Medicine, MSU
Duration: 04 April 2024 - 08 August 2024
Improved document retrieval recall by 22% in a clinical RAG system using AWS Bedrock and OpenSearch, and deployed a scalable ML backend on FastAPI/ECS.
AI Research Summer Associate Intern
AI Research Team, JP Morgan & Chase
Duration: 04 June 2023 - 26 August 2023
I worked on developing state of the art financial question answering system using Retrieval Augmented Generation paired with advanced prompting techniques.
NLP Engineer
Real World Data (RWD) Team, ConcertAI (SymphonyAI)
Duration: 06 Jan 2021 - 30 July 2022
I worked on developing NER and Relationship classification models on EHR data to better assess patients fit for clinical trials.
Data Scientist
Innovation Team, Innoplexus Consulting Pvt. Ltd.
Duration: 15 August 2017 - 05 January 2021
I worked on building NER, knowledge graph, sentiment analysis and topic extraction models on biomedical datasets.
Data Science Intern
Innovation Team, Innoplexus Consulting Pvt. Ltd.
Duration: 17 May 2017 - 20 July 2017
I worked on bulding web-scraping pipelines and developing ml methods to improve sharding in MongoDB database.
portfolio
projects
AI Voice Studio
Full-stack voice processing platform with voice cloning, multilingual TTS, and ASR transcription, using Nvidia NIM speech models. (Repo Link)
Knowledge Base Q&A Application
Desktop app for managing a personal knowledge base, with hybrid retrieval and LLM-powered, citation-grounded Q&A. (Repo Link)
Financial Entity Relationship Classification App
Extracting and classifying relationships between entities in financial texts using LangChain and LLMs. (Repo Link)
Predicting Funding Flows in Public and Private Markets (Master’s Capstone Project)
Our project aims to provide a robust framework for forecasting funding flows in public and private markets by harnessing the power of data science, time series analysis, and NLP. (Repo Link)
Concept Map Browser Extension
A browser extension which will take the selected text on a website as input and create a concept map out of it. (Repo Link)
Bio Medical Bert Named Entity Recognition
A repo for training bert based models for biomedical named entity recognition models. (Repo Link)
BioMedicalAbbvExtractor
Identifying abbreviations and their corresponding long forms from biomedical abstracts. (Repo Link)
TorchServe Based Model Finetuning Service
Model Finetuning Service using TorchServe. (Repo Link)
Dunamic Convolutional Neural Network
Paper implementation. (Repo Link)
Very Deep Convolutional Networks for Text Classification
Paper implementation. (Repo Link)
Experiments with TensorFlow and Keras
Code pieces. (Repo Link)
publications
Uncovering the Corona Virus Map Using Deep Entities and Relationship Models
Published in Arxiv Preprint, 2020
This paper focuses on extracting COVID-19 entities and relationships from a corpus of articles using a novel model. Through multi-task learning on annotated data, we train entity recognition and relationship discovery models. Employing concept masking, we prevent neural networks from functioning as associative memory, promoting contextual inference. Our findings reveal significant subnetworks, emphasize key terms, and illuminate past treatment modalities for related ailments.
Recommended citation: K Singh, P Singla, K Sarode, A Chandrakar and C Nichkawde (2020). " Uncovering the Corona Virus Map Using Deep Entities and Relationship Models" preprint ArXiv:2009.03068. 1(1). https://arxiv.org/abs/2009.03068
NLP Driven Political Analysis of Subreddits
Published in 3rd International Conference on Data Intelligence and Cognitive Informatics, 2022, Tirunelveli, India, 2022
This paper explores diverse applications, delving into model mechanics and decision-making processes. Analyzing extensive text data, we employ interpretable algorithms to extract insights. Specifically, our study focuses on extracting information from political subreddits using sentiment analysis and classification model features.
Recommended citation: Singh, K., Anne, S.V.N.S. (2023). "NLP Driven Political Analysis of Subreddits" Data Intelligence and Cognitive Informatics. Algorithms for Intelligent Systems. Springer, Singapore.. 1(2). https://link.springer.com/chapter/10.1007/978-981-19-6004-8_54
Soft Measures for Extracting Causal Collective Intelligence
Published in NLP4Science Workshop, EMNLP 2024, Miami, FL, 2024
This paper introduces soft, LLM-based measures for extracting causal collective intelligence from group discussions and text, moving beyond rigid rule-based extraction toward more flexible, nuanced causal representations.
Recommended citation: Maryam Berijanian, Spencer Dork, Kuldeep Singh (2024). "Soft Measures for Extracting Causal Collective Intelligence" NLP4Science Workshop, EMNLP 2024. https://arxiv.org/abs/2409.18911
FinQAPT: Empowering Financial Decisions with End-to-End LLM-driven Question Answering Pipeline
Published in 5th ACM International Conference on AI in Finance (ICAIF 2024), New York, NY, 2024
FinQAPT is an end-to-end pipeline that closes the gaps in real-world financial question answering, from identifying relevant documents to reasoning over structured and unstructured data using LLMs, achieving state-of-the-art results on FinQA. Recognized as Best Paper Runner-Up at ACM ICAIF 2024.
Recommended citation: Kuldeep Singh, Simerjot Kaur, Charese Smiley (2024). "FinQAPT: Empowering Financial Decisions with End-to-End LLM-driven Question Answering Pipeline" 5th ACM International Conference on AI in Finance (ICAIF 2024). Best Paper Runner-Up. https://arxiv.org/abs/2410.13959
Comparative Analysis of Multi-Agent Architectures for Entity Relationship Classification
Published in 18th International Conference on Agents and Artificial Intelligence (ICAART 2026), 2026
A comparative study of multi-agent LLM architectures for entity relationship classification, evaluating how different agent decompositions and coordination strategies affect extraction quality.
Recommended citation: Maryam Berijanian, Kuldeep Singh, Amin Sehati (2026). "Comparative Analysis of Multi-Agent Architectures for Entity Relationship Classification" 18th International Conference on Agents and Artificial Intelligence (ICAART 2026). https://arxiv.org/abs/2506.02426