Hello, I am Supriya Khadka
NLP & HCI Researcher
NLP & HCI Researcher
I am an NLP & HCI Researcher with interests in Human-AI interaction and User-centred Privacy. My goal is to build intelligent systems that are not only efficient but also trustworthy and secure, particularly in sensitive social and medical settings.
I recently completed my Master of Science by Research (MRes) at the Research Centre for Computational Science & Mathematical Modelling, Coventry University, UK. My thesis, "Automated Clinical Coding with a Hybrid Evidence-Based Pipeline: Investigating Dataset Quality, Undercoding and Coding Errors," supervised by Dr. Xiaorui Jiang and Prof. Vasile Palade, focused on optimising the clinical coding process using large and pre-trained language models. This research and my degree at Coventry were supported by the prestigious British Council Women in STEM Scholarship.
My technical foundation was built at the Institute of Engineering, Pulchowk Campus (Tribhuvan University) in Nepal, where I completed my Bachelor's in Computer Engineering. It was there, amidst a vibrant community of brilliant minds, that I first experimented with machine learning and discovered the potential of technology to solve real-world problems.
I love going on hikes, even though my body gives up almost instantly when climbing uphill! I also love buying novels, reading novels, and talking about novels in my free time. If you ask me for recommendations, these are probably the ones I will suggest. The list includes every genre; take your pick!
I also enjoy watching live theatre, movies, and series. If you want to talk about NLP, books, or movies, hit me up!
Explored evidence extraction, code prediction, and verification for automated clinical coding with LLMs and PLMs while conducting error analysis with professional coders.
Focus: Healthcare NLP, Human-AI Collaboration
Code (Paper 1) | Paper 1 (ICMLA) | Paper 2 (Preprint)
Evaluated gender bias in Nepali-English Machine Translation systems by adapting benchmarks for gender-neutral and gender-specific contexts.
Focus: Societal Impact of AI
Code & Data |
Paper (GeBNLP 2025)
An audiobook platform designed for accessibility, bridging the literacy gap using an integrated Nepali Text-to-Speech engine. Features a PDF reader and built-in audio player.
Developed a natural-sounding Nepali TTS system finetuned on a custom curated dataset. Implemented utilizing Tacotron-2 (spectrogram) and Hifi-GAN (vocoder).
Adapted IndicXlit for Nepali systems by creating a parallel corpus of ~3500 Romanized-Nepali words, capturing conversational nuances. Achieved 86% Top@5 accuracy.
An automated pipeline to generate Knowledge Graphs from business news articles, utilizing SpaCy for Entity/Relation extraction and NetworkX for visualization.
Project: Lowering the Cost of Healthcare by Clinical Coding with LLMs
Supervisors:
Dr. Xiaorui Jiang,
Prof. Vasile Palade
| Scholarships | Awarded By | Year |
|---|---|---|
| British Council Women in STEM Scholarship | British Council | 2024 |
| Ncell Excellence Scholarship - Yearwise Top Female Student (4 times winner) | Ncell | 2019-2023 |
| Full Scholarship Recipient of NAAMII Winter School of AI, 2021 | NAAMII | 2022 |
| Grace Hopper Celebration Student Scholar | AnitaB.org | 2021 |
| Women Leaders in Technology Fellowship 2020/21 | WLiT Nepal | 2020 |
While transliteration seems straightforward, the accuracy and efficiency of the process can vary depending on the ...
You have finally decided on a project idea, and are all set to take the world by storm by coding the most amazing application.