About Skills Experience Projects Research Contact Get in Touch
Lateef Olatunji
Open to Work
Available for research & collaboration

Lateef Oluwatobi
Olatunji

@Relixsx· ML Engineer· AI Researcher· Electrical Engineer

Building intelligent systems that solve real problems in healthcare,
energy, and infrastructure — from research to production.

View Projects Get in Touch
SCROLL ↓
01

EXPERIENCE

CURRENT
2022 — Present
Electrical Engineer
Transmission Company of Nigeria (TCN) · Nigeria

Planning, maintenance, and optimization of high-voltage transmission infrastructure across Nigeria's national grid. Applies data-driven diagnostics and sensor analytics to improve grid reliability and reduce downtime across critical substations.

Key Achievements
✓High-voltage transmission planning across Nigeria's national grid
✓Data-driven diagnostics and sensor analytics for grid reliability
✓Reduced critical substation downtime through predictive maintenance
Technologies Used
Power SystemsGrid OptimizationSCADAHV TransmissionFault Diagnosis
CURRENT
2023 — Present
AI Engineer
Freelance · Remote

Designs and deploys end-to-end ML pipelines for clients across medical AI, energy, and fintech. Delivers production-grade models with FastAPI backends, React dashboards, and Grad-CAM explainability modules on tight delivery cycles.

Key Achievements
✓Deployed DenseNet-121 breast cancer classifier with 97%+ accuracy
✓Built LSTM/CNN-LSTM pipeline leak detection system for oil & gas
✓Delivered production FastAPI + React dashboards across multiple domains
Technologies Used
PyTorchFastAPIReactCNN / LSTMGrad-CAMHuggingFace
2022
Electrical and Automation Engineer Intern
SanVerad Nigeria Limited · Nigeria

Supported engineering operations across electrical and automation systems. Gained hands-on experience with industrial control systems, PLC programming, and instrumentation in a production environment.

Key Achievements
✓Hands-on experience with industrial control systems and PLC programming
✓Instrumentation and sensor calibration in a production environment
✓Supported electrical system maintenance and automation workflows
Technologies Used
AutomationPLCInstrumentationElectrical SystemsControl Systems
02

PROJECTS

← SWIPE TO EXPLORE →

MedAI — Dual-Modality Breast Cancer Detection
DenseNet-121 (histopathology) + 3× EfficientNet-B4 ensemble (mammography)

Production-grade, explainable breast-cancer detection across two imaging modalities — H&E histopathology patches and full-field mammograms. PyTorch models served through a FastAPI backend with Grad-CAM heatmaps and an LLM clinical assistant, deployed live as a public web app.

◎KEY ACHIEVEMENTS
  • Dual-modality pipeline — DenseNet-121 histopathology (88.0% acc, 87.5% sensitivity on PCam) + 3-model EfficientNet-B4 ensemble for mammography (0.8443 patient-level AUC on RSNA)
  • Grad-CAM heatmaps on both modalities for explainable, clinician-facing decision support
  • LLM clinical assistant (gpt-oss-120b) with clinician / researcher / patient explanation modes
  • FastAPI REST backend — real-time + batch inference, DICOM / VOI-LUT support, BI-RADS scoring, 23-test pytest suite
  • Deployed end-to-end — Hugging Face Space backend + Truehost frontend, with automated uptime keep-alive
PyTorchDenseNet-121EfficientNet-B4FastAPIGrad-CAMLLMDICOM
Live Demo Code
🔬
0.84
Patient AUC
PipeGuard AI
LSTM / CNN-LSTM Pipeline Leak Detection

Oil & gas SCADA pipeline leak detection using LSTM and CNN-LSTM hybrid architecture. Real-time React dashboard with multi-sensor data fusion and adaptive threshold alerting.

◎KEY ACHIEVEMENTS
  • LSTM & CNN-LSTM hybrid for temporal anomaly detection
  • Multi-sensor SCADA data ingestion pipeline
  • Real-time React dashboard with alert system
  • Adaptive leak severity classification engine
LSTMCNN-LSTMSCADAReactPython
Live Demo Code
📈
<2%
False Alarm Rate
RialoLens Adventures
Gamified Blockchain Learning Platform

Full-stack gamified learning platform for the Rialo blockchain ecosystem. React + FastAPI + Neon PostgreSQL on Railway with OAuth, JWT auth, XP mechanics, and iteratively debugged quiz engine.

◎KEY ACHIEVEMENTS
  • OAuth + JWT auth with session persistence
  • PostgreSQL on Neon DB + Railway deployment
  • XP, streaks & leaderboard quiz engine
  • Full Vercel + Railway deployment pipeline
ReactFastAPIPostgreSQLJWTBlockchain
Live Demo Code
🎮
XP
Gamified Learning
OptiSub
AI Subscription Optimization Platform

Full-stack AI platform chaining Isolation Forest, KMeans, Prophet, and PuLP across anomaly detection, segmentation, forecasting, and cost optimization with blockchain integration.

◎KEY ACHIEVEMENTS
  • 4-model ML chain: anomaly → cluster → forecast → optimize
  • Prophet time-series revenue forecasting
  • PuLP linear programming cost optimization
  • Blockchain audit trail integration
ProphetKMeansPuLPReactBlockchain
Live Demo Code
📈
4×
ML Models Chained
RialoLens AI Explainer
Audience-Tailored NLP Content Engine

Streamlit NLP app using FLAN-T5 and DistilBART to generate audience-tailored blockchain concept explanations. Adjusts reading level per persona from beginner to expert.

◎KEY ACHIEVEMENTS
  • FLAN-T5 + DistilBART dual-model pipeline
  • Audience persona-aware content generation
  • Zero-shot blockchain concept explanation
  • Streamlit UI with streaming generation
FLAN-T5DistilBARTStreamlitNLPHuggingFace
Live Demo Code
🤖
NLP
Audience-Adaptive
03

RESEARCH

Breast Cancer Detection Using Convolutional Neural Networks on Histopathology Images
PUBLISHED · EKITI STATE UNIVERSITY · BIOMEDICAL AI · CNN / TRANSFER LEARNING
Investigates deep learning approaches for automated breast cancer classification using whole-slide histopathology image patches. Employs DenseNet-121 with transfer learning on the BreaKHis dataset, achieving state-of-the-art accuracy across binary and multi-class tumor subtype identification. Introduces clinical explainability via Grad-CAM heatmaps to support pathologist decision-making.
97%+
ACCURACY
CNN
ARCHITECTURE
XAI
GRAD-CAM
2
CLASS TASKS
04

STACK

Skills & Expertise

Technologies I've worked with in real-world projects and professional environments

Programming & Analytics
PythonC++ JavaScriptTypeScript SQLMATLAB PandasNumPy Scikit-LearnSciPy
Machine Learning & AI
PyTorchHuggingFace CNN / LSTMDenseNet-121 Grad-CAMProphet Deep LearningNLP Transfer Learning
AI Agents & Automation
LLM Integrations RAG Pipelines AI Automation FastAPI REST APIs Isolation Forest PuLP Optimization
Frontend & Visualization
ReactNext.js Tailwind CSSStreamlit TypeScriptRecharts MatplotlibSeaborn
Databases & Cloud
PostgreSQLNeon DB MySQLMongoDB JWT / OAuthDocker RailwayVercel
Electrical & Power Systems
Power SystemsSCADA HV TransmissionPLC AutomationInstrumentation Fault DiagnosisBlockchain
05

CONTACT

Get in Touch

Let's Build Something Intelligent

Have a project idea? Need an ML pipeline, AI solution, or power systems expertise?
I'm always excited to discuss new opportunities and challenges.

Email
relixsx@gmail.com
LinkedIn
Let's Connect
GitHub
github.com/Relixsx
Start a Conversation → Connect on LinkedIn View GitHub