Everyday Data Science: applied AI, agentic systems, machine learning, and AI in Africa, written by practitioners for people who build.
Reasoning Models Explained: What Test-Time Compute Actually Costs You
Letting a model think longer is a real capability (DeepSeek-R1 lifted AIME accuracy from 15.6% to 71.0%) and a real bill you pay in tokens and latency on every query. When it is worth paying, and when it is not.
Free Weekly Newsletter
Stay ahead in AI, ML & data science
Every week: curated AI research, agentic system breakdowns, Africa tech spotlights, and career intelligence, written by practitioners, for practitioners.
Agentic AI
All articles →Agent Accuracy Went Up. Agent Reliability Barely Moved.
Twelve metrics, four dimensions, 14 agentic models. The finding: reliability gains lag well behind capability gains. Why a single success rate hides the flaws that actually break deployments.
Ibrahim Denis Fofanah·6 min
Deep DiveThe Agent Revolution Is Here, and Most Organizations Are Not Ready
Gartner says 40%+ of agentic AI projects will be cancelled by 2027, and only ~130 of thousands of "agentic" vendors are real. Here's what the research actually shows, and what separates the teams shipping from the ones stuck in pilot purgatory.
Ibrahim Denis Fofanah·7 min·1
TutorialBuilding Multi-Agent Pipelines with LangGraph: A Practical Guide
LangChain now recommends the tool-based supervisor over create_supervisor. Here's the pattern to build, the three context decisions that decide if it works, and published benchmarks on which architecture actually costs less.
Ibrahim Denis Fofanah·6 min
ML & Data Science
All articles →Reasoning Models Explained: What Test-Time Compute Actually Costs You
Letting a model think longer is a real capability (DeepSeek-R1 lifted AIME accuracy from 15.6% to 71.0%) and a real bill you pay in tokens and latency on every query. When it is worth paying, and when it is not.
Ibrahim Denis Fofanah·7 min
ExplainerServing a Model: Endpoint, Batch, or Embedded?
Three ways to serve a model, endpoint, batch, or embedded, and the one question that decides which. Most people default to the always-on endpoint. It is the priciest and most fragile option, and usually the wrong one.
Ibrahim Denis Fofanah·7 min
TutorialBefore You Clean Your Data, Ask This One Question: Is It Actually Dirty?
A lender's income field was 23% empty, and the gap turned out to be the best feature in the dataset. Three mechanisms behind missing data, four diagnostics that tell them apart, and why cleaning is a provenance problem in disguise.
Ibrahim Denis Fofanah·12 min
Research Digest
All papers →Brain Waves to Words: What Brain2Qwerty Actually Does, and What It Doesn't
Meta's Brain2Qwerty decodes typed sentences from brain activity with no surgery, at 61% word accuracy. The catch: the scanner is a room, and the participants could type.
Ibrahim Denis Fofanah·6 min
ResearchRAG vs. Fine-Tuning: A 2026 Decision Framework for Practitioners
Stop arguing. Here's a decision tree grounded in cost, latency, and drift.
Ibrahim Denis Fofanah·8 min
arXiv Breakdown5 Papers That Explain How LLM Alignment Actually Works
RLHF, Constitutional AI, DPO, and Anthropic's Sleeper Agents result showing safety training can teach a model to hide rather than behave.
Ibrahim Denis Fofanah·6 min
Tools & Resources
📚 Agentic AI: Concepts, Architectures & Applications
Ibrahim Denis Fofanah's book, the definitive practitioner guide to building agent systems.
Get the book →Africa AI Spotlight
Building Intelligent Systems for the World's Fastest-Growing Markets
Africa is not just adopting AI, it's inventing new architectures for low-resource languages, unreliable infrastructure, and mobile-first contexts. Everyday Data Science brings you the stories nobody else is covering.
AI in Africa · Language Models
A 125M-Parameter Model for 11 Languages Is Not a Small Story
Ibrahim Denis Fofanah · 6 min read
Metric Reliability
The Broken Ruler: Why "LLMs Now Translate African Languages" Doesn't Mean What You Think
Ibrahim Denis Fofanah · 7 min read
AI in Africa · Rwanda
Rwanda Is Betting It Can Lead African AI Without Owning the Compute
Ibrahim Denis Fofanah · 7 min read
AI in Africa · Sierra Leone
Sierra Leone Is Writing Its First AI Strategy. The Minister Asked a Better Question Than Most Rich Countries Do.
Ibrahim Denis Fofanah · 7 min read
Ibrahim Denis Fofanah
Data Scientist & AI Researcher
From the Editor
Why I Built This, and Who It's For
I'm a data scientist and AI researcher at Pace University's Seidenberg School, and founder of the Rise Africa Foundation for STEM and Innovation in Sierra Leone. My research keeps circling one question: what happens to the people a system doesn't see? That question produced Artificial Frictional Unemployment, the finding that automated hiring systems reject qualified people not for lack of skill, but because of how algorithms read them (arXiv:2601.14534). And it produced the first machine learning model for crop yield prediction in Sierra Leone, where a country plans its food security largely blind, not because the technology is hard, but because nobody had built it (arXiv:2606.13959). I'm also the author of Understanding Agentic AI. I started Everyday Data Science because I got tired of reading AI writing by people who don't build things. Everything here is meant to be usable, and honest about what doesn't work, including my own results.