AI Development
Build document assistants, RAG applications, and model-powered workflows that fit a specific business task.
AI Engineer · AI DeveloperMachine Learning Engineer · B.Sc. Biomedical Technology
Oguma Eluanatein Odo · Eluan the Genius
AI & machine learning engineering.
From data to working systems.
I build retrieval systems, machine learning APIs, and data tools that people can run and inspect. I help teams turn documents, signals, and datasets into useful software—with clear evaluation and a practical path to deployment.
Approach
How systems get built
Evaluation first
Multi-model comparison and inter-patient or held-out splits—not a single cherry-picked score.
Runnable artifacts
Synthetic or documented data paths, tests, and CI so a reviewer can exercise the work.
Service boundaries
Models leave notebooks as APIs, containers, or apps with clear failure modes.
How I can help
AI, software, data & STEM services
Practical engineering and learning support for teams, founders, researchers, and educators across Nigeria and international remote clients. Explore the service that matches your problem; we agree on deliverables and success criteria before work begins.
Machine Learning
Develop and compare predictive models, then package the chosen approach for practical use.
ML Engineer · Applied MLPython Development
Create Python APIs, automation scripts, and reusable tools around your data or workflow.
Python Developer · Backend DeveloperData Science
Investigate a dataset, test useful hypotheses, and compare statistical or machine learning approaches.
Data ScientistData Analytics
Make operational data easier to understand through cleaning, metrics, charts, and decision tools.
Data Analyst · Business IntelligenceData Engineering
Build focused ingestion and transformation pipelines that prepare data for analysis or machine learning.
Data EngineerSoftware Development
Turn a focused product idea into a usable application, API, or engineering tool.
Software EngineerAI Consulting
Decide where AI can help, what data it needs, and how to evaluate a useful first version.
AI Consultant · AI SolutionsRobotics Engineering
Scope educational robotics, sensor-processing, and software automation prototypes.
Robotics Engineer · AutomationTechnical Training & STEM Education
Plan practical learning sessions in Python, AI fundamentals, data skills, and introductory robotics.
STEM Instructor · Robotics CoachSelected engineering work
Systems with evidence behind them
A selection of my public projects, chosen for implementation depth, reproducibility, and the problems they explore. Each repository explains what works and what still needs validation.
01
NeuralForge
A native inference and local-learning toolkit for CPU applications. I built a C++20 core, C API, deployment bundles, checkpoint resume, candidate evaluation, and rollback.
Evidence: participant-held-out UCI HAR evaluation, numerical comparisons with PyTorch, and x86-to-ARM64 artifact checks. The logistic regression baseline is stronger on this task; physical-device measurements remain future work.
C++20C APICMakePythonARM64
02
LocalCodeGraph
An offline repository-context toolkit for coding agents. Python AST indexing, SQLite storage, BM25 retrieval, optional dense search, and bounded dependency expansion help select code within a fixed context budget.
Evidence: budget-matched comparisons on 23 SWE-bench Lite development tasks. Changed-file hits tie BM25 at 11/23; a separate post-hoc pilot produced one test-verified BM25 repair. No graph superiority claim.
PythonSQLiteASTBM25Retrieval
03
Elite Data Intelligence
My consulting and software project brings together a browser-local CSV workbench and a bring-your-own-key AI workspace with 15 specialist roles, structured outputs, and human review.
What to explore: CSV profiling and charts, provider configuration, quality checks, and downloadable outputs. Agent workflows draft work products; client portal and booking features remain placeholders.
Next.jsTypeScriptData analyticsAI workflows
04
ECG arrhythmia research
A patient-separated ECG classification study with Random Forest and CNN comparisons, error analysis, a waveform explorer, and a FastAPI inference prototype.
Evidence: measured MIT-BIH comparisons and external/noise evaluations with documented limitations. Minority-class performance remains weak. Research and education only; not a medical device.
PyTorchscikit-learnFastAPIDockerSignal processing
Additional
Other systems
Industrial signal analytics
DSP preprocessing, predictive maintenance on synthetic sensor series, SQL calibration tracking.
Sales performance analytics
Generate, clean, and query retail sales data; export for spreadsheets or BI tools.
Predictive maintenance
Random Forest failure prediction on synthetic process features with offline metrics.
Stack
Technical tools
Languages
Python, SQL, C++20, TypeScript
ML / DL
PyTorch, scikit-learn
LLM / RAG
LangChain, FAISS, Hugging Face
Imaging
DICOM, NIfTI, Three.js
Serving
FastAPI, Docker, CI
Continued learning
Certifications & course completions
Formal coursework supporting my practical work in Python, machine learning, generative AI, agents, and data analytics.
CS50’s Introduction to Programming with Python
Harvard University
Generative AI: Working with Large Language Models
LinkedIn Learning
Artificial Intelligence Foundations: Machine Learning
LinkedIn Learning
Fundamentals of Agents — Agents Course Unit 1
Hugging Face
Intro to Deep Learning; Intro to Machine Learning
Kaggle
Learning Data Analytics
LinkedIn Learning
About
Background
I’m Oguma Eluanatein Odo, also known as Eluan the Genius—an AI/ML engineer, Biomedical Technology graduate, and founder of Elite Data Intelligence. My work connects retrieval, local inference, biomedical signals, and useful data interfaces. I care about understandable implementations, fair model comparisons, and software another person can reproduce.
I also work as a STEM Engineer and mentor and coach at World STEM Festival. Open to AI/ML engineering roles, freelance projects, and research collaborations across Nigeria and internationally through remote work. Explore my public work above or get in touch to discuss a concrete problem.
Before we build
Working together
What do you need to scope a project?
Send the problem you want to solve, who will use the result, the data you can share, and your target timeline. I’ll use those details to discuss a practical scope and acceptance criteria.
Can you improve an existing ML or RAG prototype?
Yes. We can start with its data flow, evaluation, and failure cases, then scope improvements to retrieval, model comparisons, APIs, or the user interface.
Where can I inspect your work?
Each featured project links to public source and its documentation. NeuralForge and LocalCodeGraph include measured comparisons; the ECG project includes research reports and an interactive explorer.
Let’s work together
Have a problem worth building for?
Tell me about your data, your users, and the decision or workflow you want to improve. Let’s define a useful first version.
Email me about your project →