MIRAJ BHATTARAI

machine learning · kathmandu

miraj@portfolio:~$

$ whoami

Miraj Bhattarai

Machine Learning Intern @ Techkey Technology Pvt. Ltd.

Final-year B.Sc. CSIT student in Kathmandu. I train models in Python and build the APIs and interfaces that put them in front of people — wildfire risk across Nepal, OCR on messy bank statements, sensors watching a forest.

$ get résumé $ ls projects
  • statusopen to ML & software roles
  • basedKathmandu, Nepal · UTC+5:45
  • wins3 hackathons
~/miraj.png
Miraj Bhattarai, machine learning intern and software developer based in Kathmandu, Nepal

ज्ञानं परमं बलम्knowledge is the supreme strength

About Miraj Bhattarai

about.md

I'm in the final year of a B.Sc. in Computer Science & Information Technology at Tribhuvan University, and I currently work as a Machine Learning Intern at Techkey Technology Pvt. Ltd. in Kathmandu — building and training models in Python, running experiments, and turning the ones that work into code the team can maintain.

Almost everything I've built started as a real problem, not a tutorial. A wildfire risk system for Nepal that pulls live satellite and weather data. A QR system that replaced a university's paper lab submissions. An OCR tool that reads bank statements nobody wants to type out. A Raspberry Pi in a forest listening for gunshots and smoke.

What I actually enjoy is the whole path — cleaning the data, choosing the model, arguing with the evaluation metrics, then building the API and the screen that make any of it useful to a person who will never see the notebook.

# currently going deeper on deep learning with PyTorch, and on getting models into production properly.

Experience

experience.log

Machine Learning Intern

Techkey Technology Pvt. Ltd. · Kathmandu, Nepal

Aug 2026 — present
  • INFO Build and train machine learning models in Python with Scikit-learn, TensorFlow and PyTorch — data preprocessing, feature engineering, hyperparameter tuning, and evaluation on accuracy, precision and recall.
  • INFO Research and prototype candidate approaches for assigned problems, running structured experiments and benchmarking every model against a baseline before one is taken forward.
  • INFO Document datasets, model architectures and experiment results so training runs stay reproducible and reviewable by the wider team.
  • INFO Refactor research prototypes into modular Python, working through code review with senior engineers.
  • Python
  • Scikit-learn
  • TensorFlow
  • PyTorch
  • Model evaluation

Machine learning and software projects

Four builds. Pick one to open it.

$ cat projects/agnidrishti-nepal/README.md

AgniDrishti Nepal

AI wildfire prediction & early warning system for Nepal

final year major project

problem
Wildfires spread across Nepal faster than anyone can report them, and the people at risk find out last.
approach
Custom-trained Random Forest and Naive Bayes models scoring wildfire risk from weather, elevation and satellite fire-hotspot data — with Vapour Pressure Deficit computed as an engineered feature.
build
A JWT-secured FastAPI backend with role-based access for admins and the public, live incident reporting, automated full-country risk scans, and a React front end with real-time hotspot mapping.
data
NASA FIRMS · OpenWeatherMap · Open-Elevation
  • React
  • FastAPI
  • MongoDB
  • Scikit-learn
  • Random Forest
  • Naive Bayes

Technical skills

Only what I've actually built with.

languages/
  • Python
  • JavaScript
  • C
  • C++
  • Java
machine-learning/
  • Scikit-learn
  • TensorFlow
  • PyTorch
  • OpenCV
  • Random Forest
  • Naive Bayes
data/
  • NumPy
  • Pandas
  • Matplotlib
  • Seaborn
backend/
  • FastAPI
  • Flask
  • Node.js
  • JWT auth
frontend/
  • React
  • Vite
  • HTML5
  • CSS3
  • Tailwind CSS
databases/
  • MySQL
  • SQLite
  • MongoDB
tools/
  • Git
  • GitHub
  • Linux
  • VS Code
  • Figma
  • Vercel
  • Render
  • Raspberry Pi
  • Azure
spoken/
  • Nepali — native
  • English — fluent
  • Hindi — fluent

Education and awards

education.txt

B.Sc. Computer Science & Information Technology

2023 — 2026

Tribhuvan University · Madan Bhandari Memorial College, Kathmandu

CGPA 3.0+ across all semesters, no backlogs.

Higher Secondary (NEB, +2 Science)

2020 — 2022

Liverpool International College, Kathmandu

CGPA 3.01.

awards.txt
  • 2026Winner — TechSprint Hackathon, Nepal Engineering College
  • 2025DataCamp Fellowship
  • 2025Lead Organizer — MBMC IdeaX
  • 2025Organizing Committee — Hult Prize at MBMC
  • 2024Winner — Tech Conclave
  • 2024Technical Category Winner — BPC Hackfest

Contact

contact

Open to machine learning and software engineering roles, internships and collaborations — Kathmandu or remote. The fastest way to reach me is email.

email
[email protected]
location
Budhanilkantha, Kathmandu, Nepal
$ send mail $ résumé.pdf