Hello, I'm

Satyasa Khadka

Computer Engineer | ML Practitioner | NLP Researcher

Computer engineering graduate focused on computer vision and NLP, with hands-on work in low-resource language modeling and script recognition.

Get in touch See my work
LinkedIn GitHub Email
Satyasa Khadka
SK

About

Computer Engineering graduate from Pulchowk Campus, IOE, Tribhuvan University. Graduated with an 84% aggregate on a government merit scholarship, ranked 43rd out of 15,000 entrance applicants.

My interests are in computer vision and NLP for low-resource languages and scripts. Currently working as a Software Engineer at DFW IT Partner while staying active in research.

Education

2021–2025

BE, Computer Engineering

Pulchowk Campus, IOE, Tribhuvan University

Aggregate 84% · Government Merit Scholarship · Ranked 43rd / 15,000

2018–2021

High School, Mathematics

Golden Gate International College

CGPA 3.85

Experience

Software Engineer Current

DFW IT Partner, Nepal

Apr 2025 – Present

Building and customizing ERP modules for clients using the Odoo framework, covering full-stack development and client-specific business workflows.

Executive Committee Member

LOCUS 2025, Nepal's Largest Tech Fest

Jun 2024 – Mar 2025

Core team member of LOCUS 2025. Mentored participants in AI, ML, and Data Science through the Software Fellowship. Organized "Girls to Code", a two-day event introducing AI and ML to aspiring female technologists.

Digital Literacy Outreach Lead

Children in Technology & WorldLink Internet Association

Sep 2024

Led a team of five across multiple districts in Nepal, running sessions for students on internet safety, cyber threats, and safe online practices.

Event Volunteer & Host

Hult Prize at IOE, Pulchowk Campus

Oct 2023 – Feb 2024

Hosted the Hult Prize on-campus event, facilitating panel discussions and introducing speakers.

Competitions & Awards

IOE Pulchowk Campus · Dec 2025

Best Project

Best project award for Handwritten Devanagari Character Recognition at IOE Pulchowk Campus (BELEX2052 batch)

Hult Prize at IOE · Feb 2024

Runner-Up

HackEye, a computer vision system for locating missing persons using facial recognition

Shequal All-Female Hackathon · Jul–Aug 2023

Finalist

Built a personalized learning app for children with autism at Nepal's largest all-female hackathon

Projects

Computer Vision · Deep Learning · Dataset
01

Handwritten Devanagari Character Recognition: Conjunct Consonants Dataset Creation and Benchmarking with Transfer Learning

Collected and published HDCoCo, a dataset of 50,020 handwritten Devanagari conjunct consonant samples largely absent from existing OCR datasets. Benchmarked recognition models using transfer learning. Published on Springer; dataset on IEEE Dataport.

Computer VisionTransfer LearningOCRDevanagari ScriptDataset Collection
View on GitHub ↗
Computer Vision · Deep Learning
02

Nepali License Plate Recognition

Used YOLOv8 for plate detection and a custom CNN for character recognition, optimized for Nepali fonts, local plate variations, and different lighting conditions. Presented as first author at ISBM Conference 2025.

YOLOv8TensorFlowKerasOpenCV
View on GitHub ↗
NLP
03

Nepali Language Auto-Completion

LSTM-based auto-completion model for Nepali text, addressing language modeling for a low-resource language and integrated into a typing interface.

LSTMPythonNLP
View on GitHub ↗
Web · Visualization
04

Sorting Algorithm Visualizer

Interactive web app for visualizing sorting algorithms in real time. Supports Bubble, Selection, Insertion, Merge, and Quick Sort.

JavaScriptHTML/CSS
View on GitHub ↗

Skills & Certifications

Languages

PythonCC++JavaScript

Libraries & Frameworks

PyTorchTensorFlowOpenCVPandasNumPyKeras

Areas of Focus

Computer VisionNLPMachine LearningDeep LearningTransfer Learning

Tools & Platforms

GitGitHubLinux

Certifications & Courses

5th Annual Nepal AI School NAAMII · 2024–25
Advanced Data Science & Artificial Intelligence Samsung Innovation Campus
Supervised Machine Learning: Regression & Classification Coursera

Publication

ISBM Conference 2025 · Kathmandu, Nepal · Springer

"Character Recognition of Nepali Number Plates"

Satyasa Khadka, et al.

Presented as first author. CNN-based character recognition for Nepali license plates, optimized for local fonts and plate variations.

View on Springer ↗
Springer

"Handwritten Devanagari Character Recognition: Conjunct Consonants Dataset Creation and Benchmarking with Transfer Learning"

Satyasa Khadka, et al.

Research on handwritten Devanagari conjunct consonant recognition, covering the HDCoCo dataset creation and model benchmarking with transfer learning.

View on Springer ↗
IEEE Dataport · Dataset

HDCoCo: Handwritten Devanagari Conjunct Consonants Dataset

Satyasa Khadka, et al.

50,020 handwritten samples of Devanagari conjunct consonants, a script combination largely absent from existing OCR datasets. Published on IEEE Dataport.

View on IEEE Dataport ↗