Writing

How the work was done: the decisions, the failures, and what was measured.

• Engineering a five-inch FPV drone from first principlesBuilding a five-inch FPV drone is an exercise in coupled constraints: every part you choose changes what the others have to do.
• Building a 360° panorama stitcher from a phone sweepA phone video, a pure-rotation camera model, and a stubborn question: how much of a 360° reconstruction pipeline can be made explicit instead of hidden behind a stitching API?
• Designing a PID library for real embedded controlThe PID equation is short. A controller that behaves predictably on an Arduino needs much more than three gains.
• Building an autonomous race car in twenty daysA small autonomous vehicle forced perception, control, power, mechanics, and team decisions into one unforgiving loop.
• Adapting super-resolution to thermal imageryUpscaling a thermal frame is not the same problem as enlarging an RGB photograph, especially when the result must run beside the rest of a perception stack.
• What a two-axis light tracker teaches about PID controlA small Arduino robot turns control theory into something visible: error, overshoot, noise, saturation, and settling all happen in front of you.
• Turning segmented cracks into measurable pathsA segmentation mask says which pixels belong to a crack. Inspection work often needs the next layer: an ordered line, a trace, and measurements that can be compared.
• Extracting medicine names from multilingual prescriptionsA small document-intelligence service for Arabic, English, and French prescriptions—and an example of why medical OCR needs explicit human verification.
• Designing election information for verifiabilityIn civic technology, the source and history of a fact matter as much as the interface that displays it.
• Building an adaptive motorcycle theory trainer for LebanonAn Arabic RTL study tool that keeps progress in the browser and spends practice time on the questions a learner is most likely to miss.
• Building a local-first AI support triage councilAn internal support workflow where an LLM proposes structure, deterministic rules enforce policy, and a human keeps the final say.
• Designing an offline-first personal finance desktop appREE treats personal finance as a private desktop workflow: fast entry, local storage, and enough structure to understand where money moves.
• Rebuilding my finance app around local-first syncJuno keeps the data on the device first and syncs only if you ask. Most of the work went into making that boring and safe.
• Building real-time Lebanese Sign Language translationLebanese Sign Language has a data problem before it has a model problem.
• Connecting posture estimation to a motorized deskA posture-aware desk closes a physical loop between vision, control, and motion.
• What four years of technical mentoring taught meTechnical mentoring is not solving the participant’s project for them.
• Building technology around crisis-response operationsIn crisis response, the interface is only one part of the system.
• What small upstream fixes teach about firmwareFive pull requests to Betaflight, PX4, and OpenFront, and the habit they share: make the mechanism explicit before touching the code.
• Talks, workshops, and teachingSessions I’ve given on on-device AI, version control, and engineering careers.
• Building an outbreak dashboard from public sourcesAggregating what agencies and newsrooms already publish into one map is useful only if every claim stays linked to where it came from.
• Designing a council of models for retinal screeningThree foundation models that disagree in the open are more useful to a clinic than one model that is confidently wrong.
• Why color science comes before the modelTo teach a network a photographer’s edit, the training pairs must be linear, calibrated, and exact. Most of the work is in the decode.
• Making tenant isolation the only pathFor a clinical platform, the dominant risk is one organization seeing another’s patients. The architecture should make that mistake hard to write.
• What a small vision venture taught me about scopeCo-founding a small applied computer-vision venture in Beirut was mostly a lesson in bounding engagements.
• A first mechatronic loop without a pumpAn early Arduino project that hit targets with gravity-fed water bursts, and taught that the plant changes after every action.