Muhammed Abdulrasheed | Building AI Systems

Physicist Turned Data Scientist | Built MATRA (AI thesis assistant), Chase BTC, Kenney Studio. I ship, not just train models.

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

Built a free ATS Simulator to remove the mystery of job applications.

Upload resume
Paste job description
Get match score and details
No more wondering why applications vanish into the void.

Try it: https://t.co/co31Ie9IRM

Works well for Tech jobs only for now.

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

I am working on a system that helps job applicants know whether a role is worth applying for before wasting time.

No more applying based on gut feeling, this helps you decide apply or skip with confidence.

Still in progress. You can look forward to the final product.

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

Just launched comPASS 🧭: AI-powered test analytics for Nigerian tutorial centers!

Turning test scores into actionable insights 🎯

📊 Instant student risk classification
📄 One-click professional reports

🔗 Live: https://t.co/1amKhJNgs1

#EdTech #AI #NigerianTech #Python

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

Chapters 1 & 2 in 12 minutes.

I built MATRA: an AI assistant that guides final-year students from
“I’m stuck” to a structured, submission-ready project.

Not a demo. It’s live.
This is an MVP.

Try it. Break it. Tell me what’s missing.
https://t.co/3sXj5WWQU9

#AI #EdTech

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

Just launched "Kenney Studio 🎨" : an AI art app I built that turns photos into realistic pencil sketches ✏️

Made with Stable Diffusion + ControlNet, hosted on Hugging Face.

See it in action 👇
https://t.co/liycND42E7

#AI #MachineLearning #ControlNet #Streamlit #KenneyStudio

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

Hey everyone, 👋 my latest project "Chase BTC"!

Chase BTC is a live system that predicts Bitcoin movements & delivers signals on a web app + Telegram bot 🚀

🌐 Try the app: https://t.co/ONG1Cs7CkH
📂 GitHub repo: https://t.co/bLBdGVKp3s

#ML #Bitcoin #Telegrambot #DataScience

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

This is my MVP version; I plan to upgrade with deep learning (Keras/TensorFlow) for smarter feedback once I finish my DL roadmap.

Big thanks to everyone inspiring my learning journey.
Feedback is welcome 🙌

#AI #SpeechRecognition #Streamlit #DeepLearning

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

I am excited to share my latest project: "Hey Rodea": An AI Speech Coach & Transcriber.

👉Try it live: https://t.co/mNejESWIb6

Upload or record your speech → get transcript + scores (Clarity, Confidence, Engagement) + feedback

#AI #SpeechTech #DataScience

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

But resilience was key I kept iterating, and here we are.

For v2.0, I’m planning to expand the dataset (500–1000 images per class) and push performance to the next level, I hope you are as excited as I am to see it.

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

I recently deployed my first Deep Learning project: "Little Botanist"

A Plant Classifier & Explorer built with TensorFlow (CNN)

Trained on 100 classes, optimized to run smoothly on CPU, and deployed live.

Try it here: https://t.co/TcXi1Szhb4

#DeepLearning #ML #AI

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

And I tackled the underfitting by tuning augmentations & model strategy.
- Dataset size (10k images / 100 classes) resulted in limited accuracy (~60%)
- I had to train overnight, created multiple Colab accounts and son, this project really taught me a lot.

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

This is just v1.0, and I faced quite some challenges during the development:
- Training on CPU = 7+ hours per session
- Colab usage limits forced me to get creative
- Overfitting (though, solved with data augmentation (Mixup, AugMix, CutMix) which resulted in Underfitting

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

I also built a predictive data mining model for clinical data.

From diabetes prediction to explainable association rules, the model identifies high-risk profiles & supports smarter healthcare decisions.

I rounded it by deploying via Streamlit for easy use.

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

Over the past few weeks, I’ve had the privilege of completing several data analysis and machine learning jobs: from supporting undergraduate and postgraduate research to practical business studies.

I’ll be sharing highlights from some of them here. Stay tuned!

#DataAnalytics

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

I explored the impact of IoT on SMEs in Osogbo, and the analysis showed that IoT is transforming SMEs in Osogbo.

Findings:
72% adoption → better efficiency, growth & profitability
Challenges: funding, skills, awareness.

Repo: https://t.co/vlEUxECxO2

#IoT #DataAnalysis #SMEs

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

I just completed an exciting unsupervised learning project!

I used K-Means Clustering and t-SNE to group 20 popular Nigerian foods based on their nutritional content.

Project: https://t.co/jmdgDScuhO

#MachineLearning #UnsupervisedLearning #DataScience #Python #GitHub

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

I delivered clean analysis, visual insights, chapter guidance & ready-to-present PowerPoint

Tools used: Python, Jupyter Notebook, PowerPoint

Looking for clean, reliable educational data analysis? I can help.

🔗 GitHub: https://t.co/Z62vwnWCT5

#DataAnalysis #Python #Education

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

I’ve recently completed two real-world data analysis projects, helping both undergraduate and postgraduate students make sense of their academic data.
I Explored how age at admission affects academic performance

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

Deployed with Streamlit so I can test it myself.

🔗 GitHub: https://t.co/HG5aB3Bf9e

If this is possible with KNN… I can’t wait to touch deep learning! 🚀

#MachineLearning #KNN #DataScience #Python #AccentClassifier #AI #Nigeria

[Muhammed Abdulrasheed | Building AI Systems @themrandroid]

Just built an Accent Classifier that predicts the native language (Hausa, Igbo, Yoruba) of a speaker from their English accent, using only traditional machine learning (KNN)!
I manually gathered the audio clips from YouTube and built the classifier from scratch.