Welcome! I'm Ronald Doku, Ph.D.

Senior Data Scientist & Machine Learning Engineer

I'm a data scientist and machine learning engineer who thrives at the intersection of innovative research and practical application.

These days I'm focused on clinical informatics and healthcare data, trustworthy and fair AI, LLM/RAG systems, ranking & retrieval, and rigorous experimentation. This space is where I share in-depth tutorials from the courses I've taught, along with my research and career.

For the things I build, visit Haske Labs, my applied AI research lab.

Download Resume Get in Touch
I specialize in
10+
Publications
6+
Years Experience
900+
Citations
Ronald Doku
Scroll to explore

About Me

Ronald Doku

Hello! I'm Ronald Doku. My life has been guided by a deep love of technology and the belief that it can transform our world for the better. This platform is where I share my ideas, projects, and sparks of inspiration, all centered on using technology to make a meaningful impact.

Academic Journey

I'm proud to share that I became the first person at Howard University to earn a bachelor's, master's, and Ph.D. in Computer Science. When the Ph.D. program launched in 2015, I was in the midst of my master's degree. After completing it, I took a one-year co-op to gain industry experience. Upon returning, I had the unique opportunity to make history as the department’s first triple-degree graduate. With many of my graduate-level courses already completed from my the master's program, I was able to focus heavily on research and accelerate my progress, completing the Ph.D. between 2017 and 2020. Outside of academia and work, I'm a lifelong soccer fan and captain of a recreational league team called AfroBeatz. Off the field, you'll often find me writing short stories, playing FIFA, or getting lost in a great book.

Personal Values

Family is at the heart of who I am: my wife, my mother, and my two sisters are a constant source of love and support. Their encouragement fuels my drive.

Below, you'll find a timeline of my journey in technology. Thank you for stopping by—I hope you'll join me in exploring the limitless potential of innovation, creativity, and collaboration!

Experience Timeline

Click on any timeline item to view detailed information

Afrusan Health

Founding Engineer

Oct 2026 - Present

  • Building a secure, auditable care-membership platform (Ghana pilot)
  • Designing server-side role/resource authorization & provenance-preserving clinical records
  • Implementing idempotent nurse task offer/accept/assign workflows

Howard University College of Medicine

Postdoctoral Research Fellow, Clinical Informatics

Aug 2025 - Aug 2026

  • Led reproducible research with All of Us, MIMIC-IV & NHANES
  • Flagged 1,482 of 19,470 (7.6%) "direct" LDL-C results matching calculated values
  • First-authored four peer-reviewed clinical studies

Haske Labs

Founder & Lead Applied AI Researcher

Oct 2024 - Present

  • ↑63.8% top-5 precision with LLM-free Bayesian/graph clinical search
  • 0.77→0.92 adverse-impact ratio with open-source governed-rank
  • ↓73% measured disparity with the GBP-Audit bias-correction framework

BizzySync

Machine Learning Engineer / Data Scientist

Oct 2024 - May 2025

  • Developed AI-powered partnership platform for small businesses & nonprofits
  • ↑20%+ plan approval rates via LLM feedback & evaluation pipeline
  • ↓~10% inference latency with asynchronous processing (Celery)

Oracle Advertising

Senior Data Scientist, ML/NLP & Experimentation

Aug 2020 - Jul 2024

  • Causal inference (DiD, PSM) to measure untested product launches
  • ↑15% conversions, ↓12% CAC through A/B testing
  • ↑7% ad relevance, ↑10% campaign efficiency with drift monitoring across 600+ classifiers
  • ↓25% model training time with SQL/PySpark & Airflow pipelines

ConsenSys

Software Engineer Intern

Jun 2018 - Aug 2018

  • Developed smart contracts on Ethereum blockchain
  • Integrated front-end interfaces with blockchain back-end
  • Built decentralized applications (dApps)

Howard University

Ph.D. Computer Science

Aug 2017 - May 2020

  • Specialized in Cybersecurity & Blockchain
  • Published research on decentralized data sharing
  • Achieved first BS+MS+Ph.D. in CS at Howard

Featured Tutorials

Predicting 30-Day Hospital Readmission for Diabetic Patients
Machine Learning Data Science Data Analysis
2026

Predicting 30-Day Hospital Readmission for Diabetic Patients

The design of a hands-on biomedical data science project, built for the RCMI Summer Data Science Research Training program: predicting whether a diabetic patient will be readmitted within 30 days, ...

Python pandas scikit-learn
View Details
Classification Models: Logistic Regression, k-NN, Decision Trees, and Naive Bayes
Machine Learning Data Science
2026

Classification Models: Logistic Regression, k-NN, Decision Trees, and Naive Bayes

A hands-on comparison of four classic classifiers on the Breast Cancer Wisconsin dataset using scikit-learn. It covers preprocessing, train/test splits, how each model draws its decision boundary, ...

Python scikit-learn pandas
View Details
Introduction to Machine Learning: Understanding Supervised Learning
Machine Learning Data Science
2026

Introduction to Machine Learning: Understanding Supervised Learning

A first look at machine learning that treats data as geometry: how real-world things become vectors, classification versus regression, how linear regression learns a mapping by minimizing squared e...

Python NumPy scikit-learn
View Details
Statistics Foundations: Mean, Spread, Sampling, and Your First Hypothesis Tests
Statistics Data Science Statistical Analysis
2026

Statistics Foundations: Mean, Spread, Sampling, and Your First Hypothesis Tests

A beginner-friendly tour of the statistics behind data science and A/B testing: central tendency, spread, sampling and standard error, the Central Limit Theorem, hypothesis testing with z-scores, p...

Python NumPy SciPy pandas
View Details
A/B Testing & Statistical Thinking with 2022 BRFSS Data
A/B Testing Statistics Experimentation Statistical Analysis
2026

A/B Testing & Statistical Thinking with 2022 BRFSS Data

A practical walkthrough of the statistics behind A/B testing, framed around public health interventions and the CDC's Behavioral Risk Factor Surveillance System (BRFSS). It covers sampling, spread,...

Python SciPy statsmodels
View Details
Mastering Trade-Off Analysis
Data Analysis Product Analytics A/B Testing Data Science
2026

Mastering Trade-Off Analysis

Part 5 of the product analytics series: an eight-step framework for balancing competing metrics like speed and conversion, from objectives with a floor and baselines to regression, multi-variant A/...

View Details
View All Tutorials

Skills & Expertise

ML / NLP / LLM

Python SQL PyTorch scikit-learn Transformers Embeddings RAG / Graph RAG Fine-tuning (LoRA, DPO) RLHF (PPO) Ranking & Retrieval

Evaluation & Experimentation

A/B Testing Causal Inference (PSM, DiD) Calibration Uncertainty Selective Prediction Concept-Drift Monitoring Offline Benchmarking

Data & Infrastructure

PySpark Airflow AWS Docker MLflow ETL Pipelines CI/CD

Healthcare Data

FHIR R4 EHR Data Pipelines Clinical Data Quality MIMIC-IV NIH All of Us NHANES

Research Publications

10+
Publications
900+
Citations
4
Research Areas

Research Focus

Clinical Informatics & Data Provenance

Laboratory-data provenance, ML benchmark integrity, and decision-threshold reliability with All of Us, MIMIC-IV, and NHANES

Trustworthy & Fair AI

Bias correction with performance guardrails (GBP-Audit) and constrained, fairness-aware ranking (governed-rank)

Ranking, Retrieval & Abstention

When ranked decision systems should abstain (Confidence Gate Theorem), confidence-gated routing, and graph-based retrieval

Security, Blockchain & Federated Learning

Doctoral and earlier work on big-data security, decentralized data sharing, and federated learning

Featured Publications

new

Detecting Friedewald-Substituted “Direct” LDL Cholesterol in a National Multi-Site EHR Cohort: Prevalence and Benchmarking Impact

R Doku, et al.
Journal of Applied Laboratory Medicine
2026
Scholar
new

Equation Disagreement as a Zero-Cost Uncertainty Signal for LDL-C Classification

R Doku, et al.
Journal of Applied Laboratory Medicine
2026
Scholar
new

LDL-C Equation Disagreement as a Threshold-Integrity Flag for Orthogonal Lipid Confirmation

R Doku, et al.
Journal of Clinical Lipidology
2026
Scholar
220 citations

Cybersecurity in big data era: From securing big data to data-driven security

DB Rawat, R Doku, M Garuba
IEEE Transactions on Services Computing 14 (6), 2055-2072
2019
View
154 citations

Blockchain technology: Emerging applications and use cases for secure and trustworthy smart systems

DB Rawat, V Chaudhary, R Doku
Journal of Cybersecurity and Privacy 1 (1), 4-18
2020
View
66 citations

Towards federated learning approach to determine data relevance in big data

R Doku, DB Rawat, C Liu
IEEE 20th International Conference on Information Reuse and Integration
2019
View

Recent Publications

GBP-Audit: AI Bias Intervention with Built-In Safety-First Guardrails

R Doku, et al. 2025

An AI bias-correction framework using geometric analysis, statistical testing, and performance guardrails, reducing measured disparity by up to 73% without meaningful accuracy loss in successful offline corrections. Presented as an invited talk.

IEEE CogMI 2025 View on Scholar

The Confidence Gate Theorem: When Should Ranked Decision Systems Abstain?

R Doku

Asks when ranked decision systems should abstain rather than act — the question behind my work on confidence-gated ranking and clinical workflow routing.

Ranking & Abstention View on Scholar

Mitigating data poisoning attacks on a federated learning-edge computing network

R Doku, DB Rawat 2021

Novel approach for identifying and mitigating data poisoning attacks in edge computing environments using federated learning techniques.

42 citations Read Paper

Iflbc: On the edge intelligence using federated learning blockchain network

R Doku, DB Rawat 2020

This paper proposes iFLBC edge, a framework combining blockchain and federated learning to address data quality issues in AI applications. Using a mechanism called Proof of Common Interest (PoCI), the system identifies relevant data from diverse sources, trains local models that are then aggregated and stored on blockchain. This approach enables network members to provide AI services at the edge, enhancing privacy and reducing latency while making intelligence more accessible to end-users.

48 citations Read Paper
View Complete Publication Record on Google Scholar

Haske Labs

Where I build. Haske Labs is my independent applied AI research lab: trustworthy ranking, abstention, and clinical AI, released as papers and open-source tools.

governed-rank

Open-source Python library for constrained ranking. Improved adverse-impact ratio from 0.77 to 0.92 on an offline COMPAS benchmark with accuracy guardrails. Provisional patent filed.

Ranking Fairness Open Source
View on Haske Labs

GBP-Audit

AI bias-correction framework using geometric analysis, statistical testing, and performance guardrails. Reduced measured disparity by up to 73% without meaningful accuracy loss. IEEE CogMI 2025.

Bias Auditing Guardrails IEEE CogMI
View on Haske Labs

The Confidence Gate Theorem

When should ranked decision systems abstain? Theory and practice of confidence-gating, including clinical workflow routing where 38.3% of 10,000 retrospective MIMIC-IV encounters met simulated automation criteria.

Abstention Selective Prediction Clinical AI
View on Haske Labs

Cookieless Personalization

Privacy-preserving ad and recommendation research: session-level intent, confidence-gating, and policy-aware ranking without persistent user identifiers, evaluated across three public datasets.

Privacy Recommendation Advertising
View on Haske Labs
Explore Haske Labs

Get in Touch

Have a question or want to collaborate? I'd love to hear from you.

Please enter your name.
Please enter a valid email address.
Please enter your message.