Hello, I'm
Indra Prasad Sapkota
AI Researcher & Computer Engineer
Exploring the frontiers of Deep Learning, Large Language Models, and Computer Vision. Building intelligent systems that make a difference.
Hello, I'm
AI Researcher & Computer Engineer
Exploring the frontiers of Deep Learning, Large Language Models, and Computer Vision. Building intelligent systems that make a difference.
I am a final-year Computer Engineering student with a strong focus on deep learning, large language models, and computer vision. I have hands-on experience re-implementing state-of-the-art LLM models (LLaMA-3 and Pali-Gemma) and computer vision applications along with deployment and backend technologies.
I aim to pursue a career in AI research and engineering, working on scalable and impactful intelligent systems.
Currently in the 4th year of my course with an aggregate of around 84%.
Graduated with a 3.7 GPA with A+ in Mathematics and Physics.
Explore my work across AI Research, Software Development, and Engineering
Re-implemented every model element of LLM from tokenizer to transformer model with grouped attention according to LLaMA paper from scratch. Tested inference with open source weights from Huggingface.
View ProjectRe-implemented Google Pali-Gemma vision LLM model from scratch with SigLIP encoder for image embeddings and Gemma language model for conditional generation.
View ProjectDeveloped a zero-shot voice cloning and text-to-speech model for English and Nepali using Resemble AI Chatterbox model with T3 language model and S3Gen audio model.
View ProjectBuilt a disease detection system for cauliflower plants using YOLOv8 with custom dataset. Deployed on React Native mobile app with Node.js backend, featuring real-time detection and voice-based communication.
Built traditional items detection system as part of YANTRA Hackathon using YOLOv8 with custom scraped dataset. Achieved mAP50 of ~0.9 and deployed using Flask and React.
View ProjectFull-featured real-time chat application with server creation, channels, WebSocket messaging, audio/video calls using LiveKit, and complete authentication system.
Pokhara University
June 2025Demonstrated the PlantCare cauliflower disease detection app and secured first place.
Object Detection Project
June 2024Built uTechsil object detection model and achieved second place.
Revenue Prediction Model
April 2025Built a revenue prediction model using real-world dataset in the national level hackathon.
ICES Collaboration
August 2025Teaching students in college who are starting in the field of AI.
Achieved third position in national level AI/ML hackathon where we built a revenue prediction model using real world dataset.
Annual AI School organized by NAAMI where we learned about AI technologies from industry experts and professors on topics like Graph Neural Networks, 3D Reconstruction, and Large Language Models.
Built a simple object detection model (uTechsil) during this hackathon and achieved second place.
Annual AI School organized by NAAMI where we learned about AI technologies from industry experts and professors on topics like Geometric Deep Learning, Computer Vision, and NLP.
Built the marketplace project (e-commerce web app with features from JWT authentication to Stripe payment integration) during this program and presented to the audiences.
Research papers and academic contributions
Presented at the IEEE International Conference on ICT and Photonics
A research paper on detecting diseases in cauliflower plants using YOLO object detection models for agricultural applications.
Feel free to reach out for collaborations, research opportunities, or just a chat!
Certificate image will be displayed here. Add your certificate images to the certificates folder.