Explain the pattern.
Applied mathematics gives complicated systems a language—from Fourier series to financial models and image reconstruction.
From Beirut · AI systems · applied mathematics
Models. Motion. Machines.
Leith Uwaydah is the founder of Primary Source and an AI builder at Local Service Spotlight. He turns complex ideas into visual models, then tests those models in working systems.
33.89° N / 35.50° E
LU
Leith UwaydahFrom Beirut, LB
Published workFive DOI-backed papers
Current practiceFounder, Primary Source · AI builder, LSS
Point of viewBuild clearly. Learn in public.
01 / The through-line
Find the hidden structure. Make it visible. Put it to work.
A yield curve, a basketball dribble, an image, and a marketing workflow look unrelated until you ask the same question of each: what pattern is underneath?
Leith’s published work explores the models. His work in AI explores what happens when those models become useful tools.
The story so farNew company · Primary Source
Give the truth a structure.
Primary Source organizes verifiable evidence so people and AI can retrieve, connect, and cite the truth. The operating rule is simple: proof first, claims second.
Leith is building the method in public through evidence inventory, entity structure, the Content Factory, and the Marketing Mechanic systems he is learning under Dennis Yu’s mentorship.
Read the Primary Source methodOne practice, three lenses
Curiosity with an output.
Follow the change.
Movement is data. One published study turns the rhythm of a basketball dribble into a periodic signal that can be modeled.
Build the system.
AI becomes valuable when it leaves the demo and enters a real workflow—clearly mapped, tested, documented, and improved.
Featured field note
The Marketing Mechanic.
A whiteboard walkthrough of how a marketing system fits together—part explanation, part build log, and a starting point for what comes next.
Research, 2024–2026
Five papers. One visible pattern.
The subjects move from periodic signals and quantitative finance to computer vision. The method stays consistent: make a difficult system legible.
View all publications
f(t) = Σ aₙ cos(nωt) + bₙ sin(nωt)
Signal → model → insight
The working loop
Think in public. Ship the proof.
Observe
Start with the real system, not the fashionable abstraction.
Model
Map the variables, relationships, and failure points.
Build
Turn the model into a tool someone can actually use.
Publish
Show the work, gather feedback, and make the next version better.
The channel is just getting started