Data Science & Economics at Hamilton College (Class of 2029). Co-Founder & CTO of EduCare LMS.

Rashid Mikidadi

Still learning. Always building.

Stylized illustration of Rashid smiling at the viewer, wearing headphones and a black EduCare hoodie, hands resting on a navy MacBook.

About

Story

I'm Rashid Mikidadi, a Data Science and Economics student at Hamilton College, originally from Arusha, Tanzania.

I've always been curious about technology and how it can be used to solve real problems. What excites me most is taking an idea, figuring out how to make it work, and seeing it become something useful to other people.

That curiosity led me to co-found EduCare, a learning management platform designed to make teaching and learning more accessible. When we started, I barely understood how a complete application worked. Through experimentation, mistakes, and a lot of learning, we gradually turned our idea into a functional platform being tested in a real school.

Since then, I've continued exploring AI, machine learning, and software development through projects, internships, and collaborations. Every experience reminds me how much more there is to learn.

I've learned that I don't need to know everything before I start building. Sometimes, building is how I learn.

Still learning. Always building.

Projects

Featured work

Flagship · EdTech

EduCare LMS + PAL

A multi-tenant learning platform built for schools with low-bandwidth needs, together with PAL, an AI assistant we are developing to read handwritten exams and generate grading and feedback.

Co-Founder & CTO · 2024 – Present · Source repository is private.

YES Accelerator Grand Prize (€4,000), March 2026

Next.jsTypeScriptFastAPIPostgreSQLSupabaseClaude visionVoyage embeddingsRedisRAGRailwayNetlify

Read the case study Explore EduCare

EduCare admin dashboard on a demonstration school account: at-a-glance metrics for students, teachers, and courses, with navigation for PAL, chat, users, courses, analytics, reports, and calendar.

Precision-recall and ROC curves on the 2015–2018 test set: XGBoost (PR-AUC 0.163, ROC-AUC 0.914) leads logistic regression, decision tree, random forest, and the Altman Z-Score baseline.

Machine Learning · AI4ALL Ignite

Corporate Bankruptcy Prediction

An XGBoost model trained on 78,682 firm-year records that substantially outperforms the classic Altman Z-Score, deployed as an interactive Streamlit application.

0.914
ROC-AUC · vs 0.794 Altman Z-Score: 0.163
PR-AUC · vs 0.025 Altman Z-Score baseline: 78,682
firm-year records

PythonJupyterXGBoostSHAPStreamlit

Project site GitHub

Data Science · iXperience Internship

Vendors First: Payment Risk

Researching public government award data and modeling which invoices will pay late, for a fintech serving businesses on government contracts.

84%
recall · late class, logistic regression: 0.76
ROC-AUC: 500
records · used in the modeling analysis

Pythonscikit-learnDatabricks

GitHub

One-vs-rest ROC curves from the logistic regression: dropout AUC 0.91, graduate 0.92, enrolled 0.80.

Machine Learning

Student Dropout Prediction

A multiclass logistic regression over 4,400+ student records for early identification of dropout risk.

76.8%
accuracy · multiclass, overall: 0.91
AUC · dropout class: 4,400+
student records

PythonLogistic regression

More charts ↓

Confusion matrix for the three classes (dropout, enrolled, and graduate), with most dropout and graduate cases classified correctly. — Confusion matrix across the three outcome classes.

Logistic regression coefficients for dropout risk: enrolled curricular units raise risk, while approved units and up-to-date tuition are protective. — What drives dropout risk: positive coefficients raise it, negative are protective.

GitHub

Web Development

Music Makes Mind

Music Makes Mind is a youth music initiative that uses rhythm, collaboration, and creativity to empower young people. I first built and deployed its public website with Bolt through prompt-driven development, then moved the project into a private GitHub repository, removed the Bolt-specific pieces, and now maintain it locally in VS Code, with pushes to GitHub deploying through Netlify.

PlannedA larger redesign, a database, event registrations, and secure online payments are planned, not yet shipped.

BoltVS CodeGitHubNetlify

musicmakesmind.org

1 / 4

Experience

Journey

Co-Founder & CTO · EduCare LMS
Role
2024 – Present
Leading engineering on a multi-tenant learning platform for schools with low-bandwidth needs, piloting with a school. Building PAL, which reads handwritten exams and generates grading and feedback.
educare-lms.org, EduCare LMS (opens in a new tab)
Ignite Fellow · AI4ALL
Program
Jun – Oct 2026
Machine learning fellowship. Led modeling on a corporate bankruptcy prediction project and helped deploy it.
Project site, AI4ALL (opens in a new tab)
Data Science Intern · iXperience × Vendors First
Role
May 18 – Jun 26, 2026
Worked on government payment-risk modeling with Python, scikit-learn, and Databricks; identified data leakage that would otherwise have inflated results.
GitHub, iXperience × Vendors First (opens in a new tab)
Grand Prize (€4,000) · YES Accelerator
Award
March 2026
Awarded to EduCare LMS.
B.A. in Data Science & Economics · Hamilton College
Education
Class of 2029
Double major in Data Science and Economics.

Skills

Toolbox

Technologies I have genuinely used, grouped by where they show up in my work. No percentage bars or expertise ratings: where it helps, the project is the proof.

Programming & Software Development

Python
My main language for data and machine learning work across the bankruptcy, Vendors First, and student dropout projects.
Bankruptcy case study
TypeScript
The language behind EduCare's web platform and this portfolio.
EduCare case study
C++

Data Science & Machine Learning

scikit-learn
Logistic regression and random forest modeling in my Vendors First internship and the dropout project.
Vendors First case study
XGBoost
The primary model in our bankruptcy prediction work.
SHAP
Explaining what actually drives the bankruptcy model's predictions.
pandas
Data preparation and analysis across the project notebooks.
Streamlit
The interactive demo our AI4ALL team shipped for the bankruptcy model.
Claude API, embeddings & RAG
The AI stack behind PAL, the exam-grading assistant we are developing for EduCare.
EduCare case study

Web Development & Deployment

Next.js & React
EduCare's frontend and this site.
EduCare case study
FastAPI
EduCare's backend services.
PostgreSQL & Supabase
EduCare's multi-tenant data layer.
Redis
Caching in EduCare's stack.
Tailwind CSS
Styling for EduCare and this portfolio.
Netlify & Railway
Deploying EduCare: Netlify for the web app, Railway for the backend.

Tools & Platforms

Databricks
The modeling environment for my Vendors First internship work.
Vendors First case study
Jupyter
Analysis notebooks for the modeling projects.
Git & GitHub
Version control and collaboration on every project here.

Beyond Code

Backstage

The parts of my life that don't show up in a repository.

Audiovisual production

Outside my major, my longest-running craft is live event production. I served as Head of AV and Productions at African Leadership Academy, and I now work as an Audiovisual Technician with Hamilton College's Audiovisual Services, supporting campus events that range from small gatherings to large productions.

The job changes with every event. Some days it's setting up projection and making sure every laptop and device in the room actually connects. Concert nights mean working behind an audio mixer, balancing sound live for an audience that reacts in real time. Other events call for video switching or a broadcast, running livestreams over Zoom or YouTube so people who aren't in the room still get the moment.

What I love most is the mix of troubleshooting and live responsibility. There is no second take: when a signal drops or a microphone dies, someone has to find the problem and fix it fast, usually while the event keeps moving. When the audience never notices the technical layer because everything simply worked, that's the win. Helping an event come together is genuinely one of my favorite things to do.

Music

I play violin and drums. Music has been part of my life for a long time, and it is still how I reset outside class: violin for the melodic side, drums for rhythm and energy.

It is also part of why I got involved with Music Makes Mind, the youth music initiative featured in my projects.

Creative work & storytelling

Beyond the booth and the practice room, I enjoy video production and creative storytelling.

The thread through everything I make, whether it's a video, a livestream, or a platform like EduCare, is the same: I like building things that connect technology with people.

Contact

Get in touch

I'm always glad to talk about internships, collaborations, EduCare, or an idea you think is worth building. Email is the fastest way to reach me, and my GitHub and LinkedIn show what I've been working on.

rmikidad@hamilton.edu GitHub LinkedIn