Full-Stack Developer | Machine Learning Engineer
Email: shrestharama65@gmail.com | Phone: (+977) 9861473522 GitHub: Rustam-Shrestha | LinkedIn: santoshshrestha4 Website: santoshshrestha4.com.np
Bachelor in Computer Application student at Saraswati Multiple Campus (Tribhuvan University, 2022 to Present). Builds full-stack web platforms and applied machine learning systems. Self-taught learner with a focus on production ML and scalable applications.
YouTube | LinkedIn | Pinterest | GitHub | Linktree
Live developer card scored from GitHub activity.
Languages: Python, JavaScript, TypeScript, Java, C, C#, PHP, MATLAB Frontend: React, Next.js, Tailwind CSS Backend: Node.js, Express.js, FastAPI, .NET, Prisma ORM Databases: PostgreSQL, MySQL, MongoDB, Redis Machine Learning and Data: XGBoost, LightGBM, CatBoost, scikit-learn, TensorFlow, PyTorch, Pandas, NumPy, OpenCV, SHAP DevOps and Tools: Git, GitHub, Docker, Kubernetes, Streamlit
Major project. Multi-tenant PERN and TypeScript loan platform with an ML credit-risk service built on FastAPI. The XGBoost model is trained on 300,000+ Home Credit records and serves real-time risk scores with SHAP explanations.
- Accuracy 85.01%, precision 24.69%, recall 41.75%, specificity 88.81%, F1 30.93%
- AUC-ROC 0.7630 (5-fold cross-validation, fold std 0.0045)
- Cost-weighted decision threshold 0.3748 minimizing business loss
- Stack: React, Node.js, Prisma, PostgreSQL, Redis, FastAPI, XGBoost
Expense tracking application with spending prediction.
Ingests spreadsheets and CSV files into PostgreSQL, supports CRUD operations through a clean UI, provides natural language querying, and renders lightweight data visualizations.
Full-stack biometric attendance system using PHP, FastAPI, MySQL, Haar Cascade, and Random Forest. 95% recognition accuracy with 30% lower query latency after optimization.
Streamlit NLP classifier using Random Forest and XGBoost with token features, substring ratios, and fuzzy matching. 72% accuracy.
Multi-vendor delivery platform in the style of Foodpanda, Uber Eats, and DoorDash.
MERN and WebSocket chat application supporting 100+ concurrent users with JWT authentication and Zustand state management.
PHP and MySQL platform with role-based access, cart, wishlist, and stock management.
Full-Stack MERN Developer (ERP Systems) JYRA Soft, YBC ERP | December 2025 to Present
- Developing ERP features with React.js frontend and Node.js backend
- Optimizing database queries and system performance
Full-Stack Developer Saurabh Group of IT, Sarbottam and Laxmi Steels | April 2025 to July 2025
- Built steel logistics backend in PHP and MySQL handling over 500K records
- Improved query performance by 35% and reduced API latency by 20%
- Implemented authenticated workflows in a team environment using Docker
Graphic Designer Kath Khabar (Kath Media), Full-time | November 2024 to March 2025 Lagankhel, Lalitpur
- Designed thumbnails and social posts for the Kath Khabar news outlet
- Skills: Adobe Photoshop, Creative Strategy
Bachelor in Computer Application Saraswati Multiple Campus, Tribhuvan University | 2022 to Present
Key Courses: Data Structures and Algorithms, OOP in Java, Database Management, Cloud Computing, Software Engineering, Web Technology, Operating Systems, .NET Technology, Information Security, Software Project Management, AI and Machine Learning
Self-directed learner. Completed the coursework below through independent study.
| Certificate | Issuer | Link |
|---|---|---|
| Math for Machine Learning | Udemy | View |
| Feature Engineering for ML in Python | DataCamp | View |
| Extreme Gradient Boosting with XGBoost | DataCamp | View |
| Introduction to Deep Learning with PyTorch | DataCamp | View |
| Intermediate Deep Learning with PyTorch | DataCamp | View |
| Explainable AI with SHAP and LIME | DataCamp | View |
| MLOps Fundamentals | DataCamp | View |
| Full-Stack Web Development | YouAccel | View |
| Web Designing | Global IT | View |
| Internship Certificate | Laxmi Steels Ltd. | View |