Data science & NLP
Building content-based recommendations from résumé and job-description text.
DATA SCIENCE · NLP · SYSTEMS
Data science meets
real-world systems.
I’m building on a background in networks and systems to explore data science, natural language processing, and practical machine learning.
Winston-Salem, North Carolina
Python, data analysis, and recommendation systems.
I’m interested in what data can tell us — and what we can build with it.
My path into data science builds on hands-on experience in networks and systems. Supporting college campuses taught me to troubleshoot carefully, understand real needs, and make technology useful for the people behind it.
Through my M.S. in Data Science at Carolina University and my capstone work, I’ve been exploring Python, natural language processing, and content-based recommendation systems. My job recommendation project brings those interests together: turning résumé text and job descriptions into explainable matches.
I also bring a practical foundation in multi-vendor networks, Linux administration, cloud services, and security — alongside experience mentoring interns and supporting teams.
नमस्ते · Hello
Nepali · EnglishData science in practice, supported by
a foundation in infrastructure.
Building content-based recommendations from résumé and job-description text.
Cleaning datasets, explaining model outputs, and making results easier to explore.
Designing campus networks and maintaining the servers and services behind them.
Supporting collaboration platforms, monitoring system health, and strengthening defenses.
Practical projects across data,
networks, and IT systems.
DATA SCIENCE CAPSTONE · 2026
A content-based recommender that compares résumé text with job postings, ranks relevant roles, and explains the terms and skills behind each match.
I built a Python workflow for text cleaning, TF-IDF feature extraction, and cosine-similarity ranking, with a Streamlit interface for exploring recommendations and potential skill gaps.
Job descriptions and résumés use varied language. The project explores how text-based similarity can surface relevant roles and make the reasons for a recommendation visible.
Clean and normalize the text, represent unigrams and bigrams with TF-IDF, then rank postings by cosine similarity. Shared terms, matched skills, and missing skills help explain each result.
The Streamlit app accepts PDF, DOCX, or TXT résumés, works with imported job data or supported job APIs, and offers filters, charts, and downloadable recommendation results.
The evaluation workflow supports human relevance ratings and calculates Precision@5, nDCG@10, and mean reciprocal rank. Similarity scores describe text alignment, rather than the likelihood of being hired.
A BROADER TECHNICAL FOUNDATION
Selected undergraduate projects
from Islington College.
FINAL-YEAR PROJECT · 2020–2021
Connecting physical sensors with a digital interface. My project explored Arduino-based vital-signal monitoring, LCD readings, and a web dashboard design.
I connected heart-rate, ECG, and gyroscopic sensors to an Arduino Mega and displayed readings on an LCD. I also developed the circuit design and wireframes for the dashboard, login, and signup screens.
The interim report documents sensor setup and interface development. GSM alerts, GPS, and full web integration formed part of the planned next stages. This was an academic development project; clinical accuracy was not established.
Work: circuit design · sensor setup · UI wireframesDIGITAL CRIME INVESTIGATION · COURSEWORK
A documented coursework simulation examining a phishing-led compromise, the evidence it leaves, and the safeguards that help reduce the risk.
I documented a simulated compromise using Metasploit, then examined the suspicious email, downloaded files, browser activity, and background process involved in the scenario.
The report connects the investigation to practical recommendations around suspicious attachments, security awareness, and keeping endpoint and firewall protections enabled.
Work: investigation report · evidence screenshots · recommendationsLAN & WAN · 2020
A practical network-simulation submission from my LAN & WAN coursework, part of the CCNA 3 & 4 curriculum and my foundation in networking.
The coursework deliverable was a Packet Tracer activity file (.pka), submitted for the final LAN & WAN coursework in June 2020.
This academic work sits alongside my later campus IT experience. The original simulation is an offline Packet Tracer activity.
Work: submitted Packet Tracer activityFrom supporting engineers to managing campus IT.
A foundation built through hands-on work.
Managed the infrastructure behind a connected campus, supporting the network and the people relying on it every day.
Supported the servers, cloud platforms, and team processes that kept college IT operations running.
Built an early foundation in IT operations and governance, risk, and compliance.
A foundation in infrastructure.
A perspective shaped by data.
Carolina University
Winston-Salem, North CarolinaIslington College
London Metropolitan UniversityGOOD THINGS START WITH A CONVERSATION
LinkedInHave a role, a challenge, or an idea in mind?
bidhayak2baral@gmail.com