Computer Engineering graduate from Tribhuvan University. Passionate about transforming raw data into intelligent systems. Currently exploring the frontiers of generative AI and cloud ML deployment.
I'm Pralahad Sapkota — also known as Pralad Sapkota or simply Pri — a computer engineering graduate who fell in love with the intersection of mathematics and code. My journey started with web development, but the moment I trained my first machine learning model and watched it accurately predict outcomes, I was hooked.
Today I design and deploy end-to-end AI systems — from data collection and EDA to model training, evaluation, and cloud deployment. I'm particularly interested in generative AI, computer vision, and making ML accessible through clean Streamlit interfaces.
A production-grade Streamlit app integrating multiple NLP pipelines. Features an LLM chatbot (Zephyr-7B), text summarization (BART), English to Nepali/Hindi translation (Helsinki-NLP), and GPT-2 next-word prediction — all in one interface.
End-to-end object detection pipeline for real-time football player tracking. Trained multiple YOLO variants with GPU acceleration, exported to ONNX, and integrated OpenCV for live video inference with bounding boxes and confidence scores.
Real-time spam classifier with 97.58% accuracy using Multinomial Naive Bayes and NLTK text preprocessing. Deployed on Streamlit.
XGBoost + Random Forest model achieving 99.95% accuracy and 0.874 AUPRC on 284,807 transactions. SMOTE for class imbalance.
Fine-tuned XGBoost classifier achieving ROC-AUC of 0.7891 on 5,110 patient samples. Rich visualizations with Seaborn and Plotly.
Multi-class SVM model for automatic news categorization with 15% precision improvement over baseline. TF-IDF feature engineering.
Responsive PHP web app with AJAX real-time search, reducing page load delays by 50%. Secure auth system managing 100+ book entries.
Enrolled at Sagarmatha Engineering College, Tribhuvan University. First contact with programming: C, C++, and the seeds of a passion for code.
Took a formal ML course at Sagarmatha Engineering College. Simultaneously built first web projects with PHP, HTML/CSS/JS — including Kitab-Hut and Fence-Alert.
Completed Kaggle's Python, Intro to ML, and Intermediate ML courses. Started the Andrew Ng Machine Learning Specialization on Coursera.
Enrolled in Broadway Infosys Python & Data Science program. Built 6+ ML projects and deployed live Streamlit apps. Completed Andrew Ng ML Specialization.
AWS Cloud Practitioner Essentials, AWS ML Foundations, Introducing Generative AI with AWS. Neural Networks & Deep Learning (Coursera) — Andrew Ng.
Building deeper intuition for neural networks. Exploring transformer architectures, vision models, and production ML deployment on AWS.
I'm open to full-time roles, freelance projects, and research collaborations. If you have an interesting AI or data problem, let's talk.