I’m Nikhil Adhikari — an M.Sc. Data Science, AI & Digital Business student at GISMA University of Applied Sciences in Berlin, headed for ML Engineering / MLOps / ML Reliability Engineering.
What I care about
Most ML portfolios show a model that runs once. I care about the part most tutorials skip: keeping systems working after they ship. Data validated before it enters, models watched in production, and failures turned into measurable, deployment-gating feedback.
One principle, three layers — data, model, system. My projects page walks through each one.
Tech Stack
The tools I build with, across the projects on this site:
- Languages — Python, SQL
- ML & data science — pandas, NumPy, scikit-learn, XGBoost, MLflow; drift detection (PSI · KS · Wasserstein), model evaluation
- MLOps & cloud — Docker, GCP Cloud Run, AWS, Kubernetes, FastAPI, GitHub Actions (CI/CD), Git
- Data engineering — Airflow, PySpark, dbt, PostgreSQL
- LLM & RAG — LangGraph, ChromaDB, Ollama, embeddings & vector search, retrieval evaluation (hit rate · MRR)
- Observability & practices — Prometheus, Grafana, pytest, ruff, mypy, Streamlit, Architecture Decision Records (ADRs)
The data-engineering and observability tools (Airflow, PySpark, dbt, Kubernetes, Prometheus, Grafana) are from my in-progress Berlin Transit build — I list them as what I’m actively working with, not yet as shipped work.
Background
- M.Sc. Data Science, AI & Digital Business — GISMA University of Applied Sciences, Berlin (2025–2027, in progress)
- Bachelor’s in Information Management (BIM) — Tribhuvan University, Kathmandu (CGPA 3.66)
See the full education page for coursework and details.
Experience
Front-End Development Intern — Mandala Infosys, Kathmandu · Apr–Jul 2024
- Built responsive front-end interfaces in HTML, CSS, and JavaScript — semantic navigation, landing-page layouts with optimised imagery, and form pages — in an agile team with sprint planning and code reviews.
- Elicited website requirements directly from clients and built and demoed reference builds to turn vague requests into concrete, reviewable options.
- Drafted solution options for the technical team and led feature prioritisation, walking non-technical clients through the opportunity cost of each choice so they could make informed trade-offs.
Currently
Open to a Werkstudent position or internship in Berlin, and casting a wide net across data and AI. I’m a strong fit for roles in Data Science, Machine Learning & AI Engineering, MLOps / ML Platform, Data Engineering, Analytics Engineering, and LLM / GenAI Engineering — anywhere I can contribute to real production systems and grow under senior mentorship.
Work authorization: I’m eligible to work in Germany as a Werkstudent and for internships — no sponsorship required.
If that sounds like your team — get in touch.
I learn slowly and deeply, from first principles. This site is where I keep the notes.