Available for opportunities in Nairobi, Kenya

Software engineer building intelligent products for healthcare, agriculture and the web.

I'm a software engineer and data scientist. I take a problem from raw data to a working interface, covering the data pipeline, the model, the backend and the front end, and I ship it to real users.

  • 20+ deployed applications
  • M-PESA payment integration
  • NHS-aligned clinical ML
  • 2 GitHub Arctic Code Vault repos
View featured work Download CV
  • HealthcareRisk screening tools for heart disease and diabetes
  • AgricultureCrop recommendation and animal disease support
  • Computer visionFacial emotion recognition with a CNN
  • Full-stack webReact and Firebase products for real clients
20+Deployed apps
10+ML models
4+Years experience
2Arctic Vault repos

Featured work

Six projects that show how I work: define the real problem, choose a fitting model or architecture, and deliver something a user can open in a browser.

Clinical machine learning
Stack R / Shiny
Standard NHS guidelines
Status Live

Cardiovascular Risk Detection

Problem

Early identification of heart failure and cardiovascular risk is critical for preventive care, but risk-scoring tools are often locked inside clinical software that is inaccessible for screening contexts.

Approach

Built an interactive risk-assessment tool in R Shiny, structuring inputs and thresholds around published NHS clinical guidelines so the output is interpretable by non-specialist users running early screening.

  • R
  • Shiny
  • Clinical ML
  • Healthcare
Launch live app
Clinical machine learning
Stack Python
Model Classifier
Status Live

Diabetes Risk Detection

Problem

Diabetes risk factors are well documented, but turning that knowledge into something a person can self-check before symptoms develop requires more than a static checklist.

Approach

Trained a classification model on standard diabetes risk indicators and shipped it as a Streamlit app, prioritizing fast, accessible early-detection screening over diagnostic precision.

  • Python
  • Streamlit
  • Classification
  • Healthcare
Launch live app
Production client work
Stack React
Client Real organization
Status Live

Heritage of Faith and Hope

Problem

A rehabilitation center needed a public-facing site that communicates trust and clarity to people in vulnerable circumstances, a context where confusing UX or a slow page has real cost.

Approach

Designed and built the full production site in React, prioritizing accessibility, clear information hierarchy, and an emotionally warm but professional visual tone throughout.

  • React JS
  • HTML5
  • CSS3
  • Client delivery
Visit live site
Computer vision and deep learning
Architecture CNN
Framework TensorFlow
Status Live

Real-Time Emotion Detection

Problem

Affective computing applications such as UX research, accessibility tools and sentiment systems need a lightweight way to read emotional signal from facial images without specialized hardware.

Approach

Built and trained a convolutional neural network for facial emotion classification in TensorFlow, then deployed it behind a Streamlit interface for real-time, browser-based inference.

  • Python
  • CNN
  • TensorFlow
  • Streamlit
Launch live app
Production client work
Stack React
Service pages 12+
Status Live

Iterin Immigration

Problem

A licensed Nairobi immigration consultancy needed a public-facing site that could make a wide, legally dense service catalogue, from work permits and passes to citizenship, residency and government liaison services, feel navigable and trustworthy to anxious, often first-time applicants.

Approach

Designed and built the full production site in React, structuring a dozen-plus dedicated service pages around a clear information architecture, and layering in click-to-chat WhatsApp support, a categorized enquiry form, an embedded office map and a blog to keep the consultancy reachable at every stage of a client's application.

  • React JS
  • HTML5
  • CSS3
  • Client delivery
Visit live site
Full-stack SaaS
Stack React
Backend Firebase
Status Live

SurveyForge

Problem

Teams need a fast way to build, distribute and analyze surveys without paying for a heavyweight enterprise tool, while still getting collaboration, respondent consent and exportable results out of the box.

Approach

Built a full-stack survey platform in React and Firebase with a drag-and-drop builder covering eleven question types, time-limited public share links with a respondent consent gate, real-time team collaboration with overwrite protection, and one-click export of results to PDF, Excel and Word.

  • React JS
  • Firebase
  • Firestore
  • PDF / DOCX / XLSX export
Launch live app

About me

Photo of Kelvin Muindi

Kelvin Muindi

@muindikelvin

  • Python
  • React JS
  • R / Shiny
  • ML
  • SQL
  • Streamlit
GitHub Arctic Code Vault contributor

I'm a software engineer and data scientist who builds machine learning models and turns them into real, usable applications. My applied work spans credit scoring, fraud detection, sentiment analysis, and cardiovascular risk prediction aligned with NHS clinical guidelines.

Alongside ML, I build full-stack web applications and dashboards with React JS, Streamlit and Tableau, including M-PESA payment integration for clients who need mobile-money checkout built directly into their product. I bring a human-computer interaction and design-thinking background to that work, so the systems I build are usable and aligned with how real users actually behave, not just technically correct.

I also work as a freelance technical researcher and writer, producing case-study-based academic papers in computer science, a practice that keeps my research and technical-communication skills sharp alongside the engineering work. Two of my repositories have been preserved in GitHub's Arctic Code Vault Archive Program, and I've independently shipped more than 20 deployed applications across healthcare AI, agriculture, computer vision and full-stack web development.

Skills and tools

Grouped by what I do with them, not just listed as tools.

AI and machine learning

  • Python
  • scikit-learn
  • pandas
  • NumPy
  • Random Forest
  • Logistic Regression
  • PCA
  • CNN
  • TensorFlow

Data science and analytics

  • R
  • SQL
  • Tableau
  • Streamlit
  • R Shiny
  • Feature engineering
  • Statistical analysis
  • BI reporting

Software engineering

  • React JS
  • JavaScript
  • HTML5
  • CSS3
  • PHP
  • MySQL
  • Firebase
  • M-PESA API integration
  • Git / GitHub

HCI and research

  • Human-computer interaction
  • Design thinking
  • Material Design
  • UX principles
  • Technical writing
  • Academic research

Proficiency levels

Python90%
React JS85%
R / Shiny82%
Machine learning88%
SQL80%
TensorFlow / CNN75%

All projects

The six projects above are my strongest work. Below are more machine learning, agriculture and full-stack builds, plus front-end practice projects used to sharpen UI skills.

Animal Disease Classification

Machine learning

Diagnostic support tool that classifies animal diseases from symptom inputs, for veterinary and agricultural use.

  • Python
  • scikit-learn
  • Streamlit
Launch app

Crop Growing Recommendation

Machine learning

Smart agriculture recommendation engine suggesting optimal crops from soil, rainfall and temperature data.

  • Python
  • Random Forest
  • Streamlit
Launch app

AI Search Assistant

AI / NLP

Conversational search application accepting natural language queries and returning contextually relevant results.

  • Python
  • NLP
  • Streamlit
Launch app

Dental Services Management

Healthcare IT

Full-stack dental practice management app with scheduling, patient records, billing and staff management.

  • React JS
  • Firebase
  • Full-stack
Launch app

E-Learning System

EdTech

Course delivery app with lesson management, student enrolment and progress tracking for online education.

  • React JS
  • Firebase
  • EdTech
Launch app

Real-Time Covid-19 Tracker

Data dashboard

Live global Covid-19 dashboard with interactive maps, time-series charts and country-level statistics.

  • React JS
  • REST APIs
  • Charts
Launch app

Video Conferencing App

Communication

Real-time video conferencing app with room creation and peer-to-peer video and audio via WebRTC.

  • React JS
  • WebRTC
  • Firebase
Launch app

Amazon Clone

Front-end practice

UI practice build replicating an e-commerce storefront with product browsing, cart and authentication.

  • React JS
  • Firebase
View project

Airbnb Clone

Front-end practice

UI practice build replicating a booking platform interface with listings, search and real-time data.

  • React JS
  • Firebase
View project

WhatsApp Clone

Front-end practice

UI practice build replicating real-time messaging, room creation and presence indicators.

  • React JS
  • Firebase
View project

Netflix Clone

Front-end practice

UI practice build replicating a streaming interface with category browsing and trailer previews via TMDB.

  • React JS
  • TMDB API
View project

Instagram Clone

Front-end practice

UI practice build with image upload, feed display, likes, comments and authentication.

  • React JS
  • Firebase
View project

Work experience

  1. Mar 2021 to present

    Freelance Software Engineer, Data Scientist and Technical Researcher

    Self-employed

    Design, build and deploy machine learning and full-stack applications independently, end to end. Also research and write academic, case-study-based computer science papers for clients, work that depends on the same rigor and clarity as the engineering side.

    • Independently shipped 20+ deployed web and ML applications across healthcare, agriculture and computer vision
    • Built multiple healthcare ML tools aligned with published clinical guidelines
    • Contributed two repositories to the GitHub Arctic Code Vault Archive Program
  2. Oct 2020 to Mar 2021

    Digital Content Officer (short-term consultancy)

    AVERC (Africa Vocational Education Researchers and Consultants)

    Coordinated digital learning content production, assigning tasks to subject-matter experts and tracking progress across multiple concurrent workstreams.

    • Compiled and formatted draft content, ensuring consistency across contributions
    • Managed task assignment and progress tracking using Google Forms
    • Collaborated directly with management and project administrators to keep delivery on schedule
  3. May 2019 to Aug 2019

    Computer Maintenance Attachment (intern)

    Kenya Wildlife Service (KWS) HQ, Lang'ata

    Hands-on experience in computer hardware maintenance, networking and system administration across KWS headquarters.

    • Assembled and maintained computer hardware, printers and network equipment for HQ staff
    • Configured network clients and managed printer networks
    • Installed and configured Windows XP, 7 and 10 across desktop machines

Education

BSc, Computer Science

University of Nairobi

Sep 2016 to Sep 2020

Second Class Honours

Kenya Certificate of Secondary Education (KCSE)

St. Andrew's High School

Feb 2011 to Nov 2014

Download my full CV

Experience, education, projects and references in one PDF.

Download CV (PDF)

Let's connect

I'm open to new opportunities, internships, graduate programs and collaborations in software engineering, applied ML or HCI research. If you have a relevant role or project in mind, reach out directly.

Emailkelvinmuindi10@gmail.com
Phone+254 704 230 589
LocationNairobi, Kenya
GitHubMuindiKelvin
LinkedInkelvinmuindi
Kagglekevinmuindi
X (Twitter)@smartprincekev1