Skip to content
View akinsola-data's full-sized avatar
  • Ibadan, Nigeria

Block or report akinsola-data

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
akinsola-data/README.md

Akinsola Emmanuel — Data Scientist & AI Engineer

Specializing in MLOps, Retrieval-Augmented Generation (RAG), Multilingual NLP & Quantitative Macroeconomics

================================================================================
EXECUTIVE ENGINEERING SUMMARY
================================================================================
Architecture Pillar : End-to-End AI Microservices (From Vector Store to Cloud Container)
Core Specialties    : Generative AI (RAG) | LLM-as-a-Judge | African NLP | Quant Finance
MLOps Discipline    : Docker Containerization | FastAPI | GitHub Actions CI/CD | Streamlit
Target Mandate      : Delivering zero-hallucination enterprise AI and production ETL pipelines
================================================================================

Professional Profile

I build, containerize, and deploy production-grade Generative AI systems, multilingual NLP microservices, LLM evaluation frameworks, and automated macroeconomic data engineering pipelines.

Combining deep quantitative rigor with enterprise MLOps discipline, I specialize in architecting air-gapped Vector Databases, fine-tuning classification pipelines for complex low-resource syntax (African language code-switching), deploying autonomous 'LLM-as-a-Judge' validation matrices, and translating raw, messy multi-source data into actionable strategic telemetry for global corporate boards and central bank compliance officers.


Core Architectural Stack

Pillar Engineering Technologies & Frameworks
Generative AI & NLP RAG Architecture, Local Vector Databases, LLM-as-a-Judge Pipelines, Advanced N-Gram TF-IDF Sublinear Scaling, Scikit-Learn
MLOps & Microservices Docker, FastAPI, Pydantic Schema Validation, Interactive Swagger UI, Joblib Serialization
Cloud & Data Engineering Hugging Face Spaces Cloud Load Balancers, GitHub Actions Cron Automation (CI/CD), Multi-Source ETL
Enterprise UI & Visualization Palantir Foundry UI, Streamlit Chat Architecture, Dual-Axis Plotting, Multiplicative Time-Series Decomposition
Domain Intelligence Numerical Contradiction Testing, Nigerian Commercial Banking Law (BOFIA 2020), CITA Tax Structures, Central Bank FX Reserve Regimes

Production Portfolio & Live Cloud Architecture

1. EvalScout AI: LLM-as-a-Judge & Hallucination Benchmark Engine

  • Live Cloud App: Launch EvalScout AI on Hugging Face Spaces
  • GitHub Repository: akinsola-data/eval-scout-ai
  • System Architecture: MLOps · LLM-as-a-Judge · Hallucination Detection · Palantir UI · Docker
  • Value Delivered: An enterprise MLOps evaluation framework that utilizes an 'LLM-as-a-Judge' pipeline to autonomously audit conversational AI models for hallucination and data groundedness. Enforces local TF-IDF cosine similarity matrices paired with exact numerical entity contradiction checking (re.findall), achieving a 100% Hallucination Catch Rate across 200 high-stakes financial and legal benchmarks.

2. AfriJuris RAG: AI Legal Assistant for Nigerian Commercial Law & CBN Regulations

  • Live Cloud App: Launch AfriJuris RAG on Hugging Face Spaces
  • GitHub Repository: akinsola-data/afri-juris-rag
  • System Architecture: Generative AI · RAG Engine · Dense Vector Database · Streamlit Chat UI · Docker MLOps
  • Value Delivered: An autonomous legal compliance platform that indexes authoritative extracts across 8 regulatory pillars (BOFIA 2020, CBN FX guidelines, CITA 2023, NDPA 2023). Implements rigid cosine similarity thresholds (> 0.05) to provide verifiable statutory citations with zero hallucination.

3. NaijaFinProtect: Multilingual Financial Fraud & Phishing Detection API

  • Live Cloud App: Launch NaijaFinProtect on Hugging Face Spaces
  • GitHub Repository: akinsola-data/naija-fin-protect
  • System Architecture: MLOps · Docker · FastAPI Backend · Streamlit Corporate Dashboard · African Language NLP
  • Value Delivered: An enterprise microservice designed for the Pan-African fintech ecosystem. Extracts high-variance TF-IDF bigrams (ngram_range=(1, 2)) to classify phishing and loan shark extortion across code-switched Pidgin, Yoruba, Igbo, Hausa, and English, achieving 100% out-of-sample accuracy.

4. Naira in Freefall: A Quantitative Analysis of Nigeria's Currency Crisis (2015–2024)

  • GitHub Repository: akinsola-data/naira-currency-analysis
  • System Architecture: Automated CI/CD (GitHub Actions) · Multi-Source ETL · Time-Series Decomposition · Ridge Regression
  • Value Delivered: A deep macroeconomic investigation merging Central Bank of Nigeria (CBN) policy dates, OPEC oil prices, and World Bank inflation proxies. Successfully proves "The Great Decoupling" (post-2021 currency detachment from oil prices) and features an automated monthly ETL cron job (0 6 1 * *) that dynamically updates the repository's hero charts.

5. Breast Cancer Prediction: Medical AI & Feature Importance

  • GitHub Repository: akinsola-data/breast-cancer-prediction
  • System Architecture: Medical Tabular Analytics · Scikit-Learn Classification · Model Telemetry
  • Value Delivered: An optimized statistical classification engine trained on highly sensitive oncology data. Demonstrates strict feature importance tracking to ensure AI explainability in clinical environments, achieving a robust 96.49% testing accuracy.

Value Proposition for Global Employers

I do not just build machine learning models in isolated Jupyter Notebooks. I deliver the complete engineering lifecycle required for modern enterprise AI:

  1. Air-Gapped Compliance & AI Safety: I ensure production models operate with strict hallucination guardrails, LLM-as-a-judge validation, and Pydantic schema enforcement to protect corporate liability.
  2. Autonomous Automation: I build self-maintaining cloud ETL pipelines that eliminate manual data engineering overhead.
  3. Executive Communication: I translate dense statistical metrics (RMSE, Cosine Distance, Attention Weights) into highly legible corporate strategies for financial risk management and regulatory compliance.

Open to global remote AI engineering roles, quantitative consulting contracts, and senior data science mandates.

Pinned Loading

  1. naija-fin-protect naija-fin-protect Public

    An enterprise MLOps multilingual NLP microservice deployed via Docker & FastAPI. Detects financial phishing, loan shark harassment, and banking fraud across African code-switched languages (Pidgin,…

    Python

  2. naira-currency-analysis naira-currency-analysis Public

    A quantitative data science investigation of Nigeria's currency crisis (2015–2024). Features multi-source macroeconomic data wrangling, time-series decomposition, structural break analysis, and mac…

    Python

  3. afri-juris-rag afri-juris-rag Public

    An enterprise Retrieval-Augmented Generation (RAG) microservice deployed via Docker & Streamlit. Instantly answers complex corporate governance queries regarding Central Bank of Nigeria (CBN) regul…

    Python 1

  4. ghost-agent-scanner ghost-agent-scanner Public

    An enterprise MLOps auditing platform deployed via Docker & Streamlit. Maps active AI agents across corporate tech stacks, identifies unmonitored "Ghost Agents" wasting API budget, and synthesizes …

    Python 1

  5. breast-cancer-prediction breast-cancer-prediction Public

    Machine learning classification project predicting breast cancer diagnosis using Logistic Regression, Decision Tree, and Random Forest.

    Jupyter Notebook

  6. eval-scout-ai eval-scout-ai Public

    An enterprise MLOps evaluation framework deployed via Docker & Streamlit. Utilizes an 'LLM-as-a-Judge' pipeline to autonomously audit conversational AI models for hallucination, numerical contradic…

    Python