Hiper Abasto Royal
A production wholesale storefront built around how the retailer actually sells. I designed the Spanish catalogue, live-offers flow, responsive discovery, and order-building path, then deployed it on AWS.
Ali Salman · Software builder
AI systems, web products, and the infrastructure behind them—designed, built, and shipped from one end to the other.
Now Building AI agents at Edgebot
Based Beirut, Lebanon
Open to useful, ambitious work.
Selected work
Products and AI systems with a clear job, shipped beyond the prototype.
A production wholesale storefront built around how the retailer actually sells. I designed the Spanish catalogue, live-offers flow, responsive discovery, and order-building path, then deployed it on AWS.
A dermocosmetics storefront that turns a dense catalogue into guided discovery. Shoppers move from a skin concern to relevant products and a practical routine without digging through irrelevant inventory.
An AI pipeline that reads construction documents and produces a reviewable Bill of Quantities workbook. It replaces hours of manual extraction with a traceable workflow estimators can check and edit.
An operations system for a Lebanese bakery-café chain. It turns nine messy report exports into demand forecasts, staffing plans, product pairings, and answers grounded in the source data.
A deployment planner for machine-learning workloads. It estimates compute, memory, and latency needs, then recommends practical hardware instead of leaving teams to size infrastructure by trial and error.
More builds
Smaller systems, same focus.
Multi-camera shopper tracking and store-level movement analytics.
Parallel Google Maps collection across a compute cluster.
Grounded course answers with measured retrieval quality.
Experience
Research and engineering work across AI, automation, and production systems.
Jun 2026 — Present
Building research-agent benchmarks and multi-agent workflows for research automation.
Apr 2026 — Present
Developing safety filters for more robust vision-language-action systems.
Nov 2025 — Feb 2026
Built and deployed the document-to-BOQ automation pipeline featured above.
Oct 2025 — Nov 2025
Integrated AI features into a production service under latency and reliability constraints.
Aug 2024 — May 2025
Implemented robust estimators for heavy-tailed data and built Python analysis tools.
2024
Led C++ and Python tutorials, supported labs, and reviewed student code.
About
I like owning the full path: shaping the product, building the interface, wiring the backend, and putting it online. AI is part of the toolkit—not the pitch.
Computer Science and Engineering at AUB, with a minor in Mathematics. I also organize AIcademy bootcamps and support workshops and hackathons with AUB’s IEEE Student Branch.
LanguagesPython, C++, Java
AI / MLPyTorch, LangGraph, LlamaIndex, RAGAS
SystemsFastAPI, Temporal, PostgreSQL, Redis, Docker
CloudAWS, CloudFront, S3, Cloud Run