Welcome to my Portfolio

Howdy, I'm John Kangethe

Explicitly a CS Researcher and Software Engineer specializing in Machine Learning, Deep Learning, Information Retrieval, Data Analysis, and AI.

Portrait of John Kangethe
Profile // 01

About Me

Origin Story

I am John Kangethe, preferably Sean, a computer scientist and software engineer on a simple mission: build intelligent systems that matter. What began as raw curiosity about how machines learn has grown into a journey through full stack engineering, large scale data systems, and applied AI research. Today that journey runs through a Master of Science in Computer Science with an Artificial Intelligence specialization, where I train models, mine massive datasets, and turn research papers into working code.

My drive lives in the space where theory meets production. I love taking an idea from a whiteboard sketch all the way to a deployed system that real people rely on. Rigorous evaluation, clean architecture, and interpretable results are the standard I hold every project to, whether it is a research prototype or a system serving users at scale.

Education

  1. M.S. in Computer Science (Artificial Intelligence)

    University of South DakotaSD, USA

    2025 — Present

  2. Certificate in Software Engineering

    Flatiron SchoolNY, USA

    2023

  3. Certificate in Web Design

    BitDegree

    2023

  4. Bachelor of Technology in Software Engineering

    Technical University

    2022

  5. Diploma in Computer Science

    PAC University

    2021

  6. Certificate in Network Fundamentals

    Cisco (CCNA Intro)

    2021
Capability Matrix // 02

Skills

Research-driven and production-ready. These reflect consistent performance on real projects, code reviews, and measurable outcomes.

Machine Learning Engineering
AI / ML
Deep Learning & LLMs
AI / ML
Natural Language Processing
AI / ML
Generative AI & RAG Systems
AI / ML
Information Retrieval
AI / ML
MLOps & Model Serving
AI / ML
Data Engineering & Pipelines
Data
Big Data & Distributed Computing
Data
Databases (SQL / NoSQL)
Data
Data Structures & Algorithms
Engineering
Software Engineering
Engineering
Full Stack Development
Engineering
API Design & Microservices
Engineering
System Design & Architecture
Systems
Cloud Architecture (AWS / GCP)
Systems
DevOps & CI/CD
Systems
Arsenal // 03

Tools & Languages

The instruments I use to turn research ideas into reliable systems.

PyTorchpro
TensorFlow
Hugging Facepro
LangChain
OpenAI API
Ollama
MLflow
Weights & Biases
Ray
ONNX
NVIDIA CUDA
OpenCV
Apache Spark
Apache Kafka
Airflow
dbt
Snowflake
Databricks
Pythonpro
scikit-learn
Pandas
NumPy
SciPy
Jupyter
Google Colab
Keras
Mission Log // 04

Works

Graduate level systems and research engineering. Selected work that shows research-grade thinking and production craft.

AI Systems

GridSintel: Power Grid Intelligence Platform

PythonPyTorchGNNsKafkaTime Series

A Graph Neural Network-based AI platform for power grid telemetry. Detects anomalies in real time sensor streams, forecasts load with deep learning, and flags stability risks before they cascade. Built around a streaming architecture with model monitoring baked in.

Research

AeroSearch: Aviation Accident Narrative Retrieval

PythonBM25StreamlitNLP

An intelligent search engine that mines aviation accident narratives from the NTSB database. Features Boolean, proximity, and BM25 retrieval models for precise query matching, built with a modular NLP pipeline, for research insights in safety analytics.

AI

StudyBuddy: AI Study Assistant

Vue 3VuetifyWeb WorkersNLP

An in-browser AI study assistant that turns messy notes or PDFs into structured revision materials in minutes. Extracts the most important ideas, then builds flashcards and short quizzes to reinforce recall. Fully client-side for privacy and speed.

Bioinformatics

Gomics: Genomics Simulation Tool

PythonBiopythonNumPyBioinformatics

A simulation toolkit for genomics research. Generates synthetic sequence data, models mutation and alignment processes, and benchmarks bioinformatics pipelines against ground truth, giving researchers a controlled sandbox before touching costly real datasets.

Systems

IoT Transmission Protocol Simulation

PythonSimPyMQTT / CoAPNetworking

A discrete event simulation of low power IoT transmission protocols. Models MQTT and CoAP behavior under packet loss, latency jitter, and constrained bandwidth to quantify trade-offs between reliability and energy budget across network topologies.

Scientific Computing

Physics Protocols Lab

PythonNumPyMatplotlibScientific Computing

A computational physics workbench implementing numerical protocols for classical and statistical mechanics. Runs N-body integrations, Monte Carlo sampling, and field visualizations with reproducible experiment configs and publication quality plots.

Research Archive // 05

Publications

Papers and working manuscripts. Read them right here without leaving the page.

Aviation Accident Narrative Retrieval with Hybrid IR Models

J. N. KangetheWorking paper, University of South Dakota2026

A study of Boolean, proximity, and BM25 retrieval over NTSB aviation accident narratives, with an evaluation of ranking quality for safety-analytics queries.

Open Channel // 06

Get in Touch

Research collaboration, engineering roles, or a good idea worth building. The channel is open.