01SEE
Make the truth visible.
Expose the data, evidence, constraints and unknowns before optimizing around them.
104 matches50k Monte Carlo
OpportunityOSTruth constrains action
01Discover02Truth-lock03Prepare04Monitor05Learn
Mohammed Ehab ElNomany
I turn complex data and business problems into reliable systems, models, analytics and products.
Credibility, compressed.
Full backgroundSelected work · 01
Three ways into the work. One standard.
The homepage keeps only three flagship systems in view. Each one carries a different kind of proof and opens into a full case study when more detail is useful.
Browse all six projectsPlot / evolvepublic_code_renderable
104 matches forecast50,000-iteration Monte Carlopost-event evaluation
Presaira
A full-tournament forecasting system that forecast all 104 matches of the 2026 World Cup, with reproducible runs, model evaluation and 50,000-iteration Monte Carlo simulation.
Data science · forecasting · calibration & evaluation · Monte Carlo · reproducibility
01Discover02Truth-lock03Prepare04Monitor05Learn
OpportunityOS
A truth-locked opportunity workflow built around governed autonomous agents, provenance and controlled generation.
Public-safe architecture · truth and provenance constrain actionBrowse to fulfillment.
01Browse02Product03Cart04Fulfillment
Ghareeb Oglu Commerce
A commerce platform planned, architected and built from scratch in code — spanning storefront, backend, payments, fulfillment and deployment.
Public-safe commerce flow · protected brand imagery remains unpublishedHow I work · 02
I like the messy part.
The part before the dashboard, model or product — when the question is fuzzy, the data is fragmented and everyone is using a different definition.
I do not start with tools. I make the problem legible, reduce ambiguity, then build the smallest reliable system that can carry the job.
INPUTmessy reality
SOURCE_01fragmented data
SOURCE_02conflicting definitions
QUESTIONwhat actually matters?
BOUNDARYevidence is incomplete
02REDUCE
Reduce ambiguity.
Turn fuzzy business questions into explicit mappings, models, decisions and interfaces.
03BUILD
Carry the job.
Build the smallest reliable system that can survive real use and real decisions.
OUTPUTreliable system
MOVEvalidated migration
EXPLAINdecision-ready analytics
PREDICTevaluated model
GOVERNAI workflow with boundaries
SHIPusable product
decision-ready
Thinking · 03
What the work taught me.
Two evidence-backed notes. The full trail stays in the essays and underlying case studies.
All writingWhen should you trust a probabilistic forecast?
A practical trust test for probabilistic forecasts: proper scoring, calibration, complete coverage, leakage-resistant evaluation, reproducibility and published failure modes.
Read the field noteFORECAST / CALIBRATION104 / 104
Proper scoringCoverageReliability
Why accuracy alone is not enough for oil-spill detection
A metric-design case study from Sentinel-1 SAR segmentation: why rare oil pixels, look-alikes and deployment domain gaps make overall accuracy a weak headline measure.
Read the field noteSAR / SEGMENTATIONOil IoU 0.566
Recall 0.7645 classesSentinel-1
Opportunity · 04
Choose the right conversation.
Role, project or system problem — the fastest route is the one with the right context attached.
01
HIRING / ROLE
Hiring for a technical, data or product role?
Start with the work, then reach me directly with the role and the problem space.
02
PROJECT / SYSTEM
Have a system or product problem worth solving?
Migration, analytics, ML/AI or product delivery — use the service context to see what is supported by public proof and what stays experience-backed.