01
Backend & APIs
Go services under sustained load. Concurrent fan-out ingestion, schema normalisation, predictable memory, no external job queue.
GoRESTGoroutinesDocker
OPEN TO FULL-TIME ENGINEERING ROLES — REMOTE
FULL-STACK
ENGINEER.
I build production systems end to end — data pipelines, backends, and the interfaces on top of them. Go, Python, TypeScript. Currently building an AI analytics platform solo, from ingestion through UI.
/ 01 — SELECTED WORK
THE WORK.
PRODUCTION SYSTEMS
& ENGINEERING WORK.
/ LAB — SELF-CONTAINED EXPERIMENTS
BUILT TO LEARN SOMETHING SPECIFIC.
/ 02 — ABOUT
PRODUCTION
SYSTEMS.
END TO END.
I'm Ahmad Chabayta, a full-stack engineer based in Beirut. I build systems that run in production and get used — not prototypes that demo well and fall over under real load.
Right now I'm the sole engineer on an AI analytics platform: a Go ingestion and API layer, a Python ML and embedding layer kept off the request path, and a Next.js interface on top. In production since 2025, used for recurring operational decisions. I own every layer of it — schema design, concurrency, inference cost, and what the user actually sees when the model isn't confident.
That last part is the work I care most about. The hard problems in a data product aren't in the happy path — they're in latency budgets, staleness tradeoffs, and designing honestly for uncertainty. A tool that confidently returns a wrong number is worse than one that returns nothing.
/ EXPERIENCE
2025 — PRESENT
Full-Stack Engineer
Sole engineer on a production AI analytics platform. Go ingestion and API layer, Python ML and embedding pipeline, Next.js interface. Architecture through deployment.
3 YEARS PRIOR
Graphic Designer
Visual design and brand work before moving into engineering. It's why my interfaces don't need a designer to fix them.
/ 03 — DISCIPLINES
CAPABILITIES.
SIX OVERLAPPING DOMAINS.
ONE PERSON. ONE STACK.
02
Data Pipelines
Multi-source ingestion on independent schedules. Heterogeneous schemas normalised into one internal representation.
ETLSchedulingPostgresPython
03
ML & Embeddings
Feature extraction, clustering, embedding generation. Kept off the request path so inference cost and p99 stay bounded.
PythonpgvectorEmbeddingsClustering
04
Databases & Search
Schema design, query performance, semantic search. Postgres over a second stateful system until the data volume earns it.
PostgreSQLpgvectorIndexingSQL
05
Frontend Engineering
Next.js server components for data-heavy views, streaming results, client code only where interaction demands it.
Next.jsReact 19TypeScriptTanStack
06
Interface & Systems
Design systems, accessibility, and motion — plus WebGL and Canvas when a product genuinely needs them.
TailwindRadixframer-motionWebGL
Stats
/ 013SERVICES IN PRODUCTION
/ 023LANGUAGES IN THE STACK
/ 031ENGINEER — INGESTION TO UI
/ 04DAYS → MINSREPORTING TURNAROUND
/ 04 — LIVE / INTERACTIVE
EXPERIMENTS.
DRAG. TAP. PLAY.
EVERY DEMO RUNS LIVE IN YOUR BROWSER.
DRAG · TAP TO BURST
CANVAS PHYSICS
Particle Gravity
120 bodies orbiting your cursor via inverse-square attraction. Tap to detonate an outward burst.
OPENDRAG · X=FREQ · Y=AMP
GENERATIVE VISUAL
Wave Synth
Three interfering waves shaped by drag. X controls frequency, Y controls amplitude.
OPENBREAKOUTDRAG TO PLAY
MICRO GAME
Breakout, 80 lines.
A complete breakout game written in vanilla canvas. Drag to play.
OPEN/ 05 — CONTACT
LET'S TALK.
I'm looking for a full-time engineering role — backend, data, or full-stack — working remotely, or relocating for the right team. Open to contract work too. Email lands fastest; WhatsApp also works.