Research Engineer

Shirshak
Acharya

Computer Vision · NLP · Data Scientist · ML Engineer

Building intelligent systems at the intersection of Medical imaging, deep learning, and scalable engineering.

Get in touch ↗ GitHub ↗ LinkedIn ↗

Scroll

🎓 M.Sc. Data Science · 🏫 TU Dortmund University · 📍 Dortmund, Germany

01 / About

The person behind
the Research

Specialized in Computer Vision, Natural Language Processing, 3D Point Clouds and Medical Imaging. Currently an M.Sc. Data Science candidate at TU Dortmund University with a strong engineering foundation (B.E. Cum Laude, 80.66%, IOE Thapathali). Experienced in leveraging frameworks like PyTorch, Lightning, Hydra, Docker, MONAI, and FastAPI to build:
i.) 3D Tooth Point Cloud Reconstruction for Intra Oral Scans, and
ii.) Semantic Segmentation for clinical CT/MRI scans
Passionate about bridging the gap between advanced statistical learning and scalable engineering solutions.

02 / Education

Academic
background

2026 — 2028
In Progress

M.Sc. Data Science

TU Dortmund University · Dortmund, Germany

Learning in Robotics · Industrial Data Science · AI · Deep Learning · Advanced Statistical Learning · ML Paradigms for Complex Data · Statistical Theory · Big Data · Data Science in Practice

2019 — 2024
Completed

B.E. Electronics, Communication & Information Engineering

IOE Thapathali Campus, TU · Kathmandu, Nepal

Cum Laude80.66%

Data Mining · Big Data · DSA · Embedded Systems · DBMS · Computer Networks · OS · AI · Linear Algebra · Calculus · Probability & Statistics

2017 — 2019
Completed

High School — Physics & Mathematics

Liverpool Int'l Secondary School · Kathmandu, Nepal

Cum LaudeGPA 3.65 / 4.0

03 / Work

Professional
experience

Mar 2025 — Mar 2026

Full-time

Foundational Research Engineer

Accelerated Komputing (AnK)

Developed latent space representation for clinical 3D dental scans using DGCNN with Contrastive Learning to automatically cluster similar tooth structures
Deployed a REST API with FastAPI: 3D mesh uploads → real-time DGCNN feature extraction → nearest neighbor matching in latent space
Generated synthetic training dataset via geometric deformations on healthy 3D teeth for crown reconstruction ground truth
Implemented MICCAI SOTA DMC framework for crown generation for INHOUSE CUSTOMER DATA using DGCNN feature extractor, transformer Encoder Decoder, and Differentiable Poisson Surface Reconstruction (DPSR)

Aug 2024 — Feb 2025

Research

Research Assistant

NAAMII — Nepal Applied Mathematics & Informatics Institute

MICCAI "PENGWIN" Pelvic Bone Fragment Segmentation Challenge using nnUNet architecture
Retina Disease Classification for Glaucoma and Age-Related Macular Degeneration (AMD)
Medical image segmentation on 3D volumes using MONAI, PyTorch Lightning, Hydra: brain tumor MRI, liver tumor CT, chest X-rays
Project Based Learning Training for BP Eye Hospital Technical Assistants for Eye Disease Classification from Fundus Images

MICCAI "PENGWIN" Pelvic Bone Fragment Segmentation Challenge using nnUNet architecture ↗ Code ↗ Project Based Learning Training for BP Eye Hospital Technical Assistants for Eye Disease Classification from Fundus Images ↗ Retinal Disease Screening Model For Nepali Eye Patients (Collaboration with BP Eye Hospital) ↗ TECHNICAL WORKSHOP For BP Eye Hospital Technical Assistants: AI Foundations, Relevant Metrics & Handling Imbalance ↗ Paper Presentation Session: EGDCL for Unbiased Glaucoma Diagnosis (Literature Review) ↗ BP Eye Web App Development Inference ↗ Docker Inference for MICCAI Submission ↗ Paper Presentation Session: Graph Convolutional Networks (Literature Review) ↗ Paper Presentation Session: Graph Attention Networks (Literature Review) ↗

Oct 2023 — Jul 2024

Internship

Research Intern

NAAMII

3D MRI Brain Tumor Segmentation (BRATS), Liver Tumor and Spleen CT segmentation using MONAI
2D Lung Segmentation from Chest X-Ray; Lung Infection Diagnosis from X-Ray using CNNs
Explored AFTer-UNet architecture for 3D segmentation tasks

UNet Anatomical Segmentation Model Using Lightning Hydra Framework ↗ 3D Semantic Segmentation with Brain MRI Scans ↗ 3D Brain MRI's Code: Using PyTorch Library ↗ 3D Brain MRI's Code: Using Monai Framework ↗ 3D Liver Tumor Segmentation using MONAI Framework ↗ 3D Spleen Segmentation with Monai & Pytorch Lightning ↗ Explored AFTer-UNet architecture for 3D segmentation (Still Pending) ↗ Paper Presentation Session: Semi VIT (Literature Review) - Version 1 ↗ Version 2 ↗

Jan 2023 — Dec 2023

Fellowship

AI Fellow

Fusemachines

Hands-on ML: Regression, Decision Trees, Ensembles, Clustering, Dimensionality Reduction, Time Series, RL
Deep Learning: CNN, RNN, LSTM, GRU, Attention, Transformers, VAEs, GANs, Deep RL
Projects: Semantic Segmentation with Brain MRI Scans, Sentiment Analysis (Naive Bayes), Visual QA (BERT & ViT), Music Genre Classification

Fusemachines CV Project: Lung Semantic SegmentationSegmentation Of 2D X-Ray ↗ Fusemachines CV Project 2's Code ↗ Visual Question Answering ↗ Sentiment Analysis with Naive Bayes Classifier ↗ Music Genre Classification ↗ GritFeat Hackathon Jan 2023: Lung Disease Classification ↗ GritFeat 2023: Lung Disease's Code ↗

Jul 2023 — Aug 2023

Part-time

Junior ML Engineer

Omdena — Nepali Chapter

Contributor: "Developing a Natural Language Toolkit for the Nepali Language"
NLP: Tokenization, Stemming, Lemmatization, zero-shot and few-shot text classification in Nepali

(Private Repo) Contributor: "Developing a Natural Language Toolkit for the Nepali Language" ↗ Certificate ↗

04 / Skills