Ghassan Alhamoud · Independent Engineering Practice
Senior Software Engineer — AI Agent Enabler, Software Architecture
I build and document agent systems in public: Tamoz, ALMS, and the ScaleShop lab, backed by 15+ years of distributed-systems practice.
Every system on this site is public, documented, and inspectable. Code, decisions, and limitations included.
The argument starts with shipped work: three public systems you can run, inspect, and trace back to their reasoning.
An evidence-driven workshop where engineers diagnose bottlenecks from telemetry and choose the smallest sufficient architecture change.
Twelve playable incident scenarios with modeled outcomes. Live and open source at lab.ghassan-alhamoud.com, with full reasoning in the repository.
Inspect the systemA Ruby-native durable agent framework for checkpointed, interruptible, observable, and evaluation-governed AI workflows.
Deterministic graph execution, SQLite durability, a RubyLLM agent runtime, and an evaluation harness are inspectable in the public repository.
Inspect the systemA self-hosted MCP server that gives autonomous agents shared operational memory without turning into an orchestration framework.
Agents register, publish learnings, and pull operational protocols by tag, and they keep working when ALMS itself is offline.
Inspect the systemApplication systems and experiments, including a film-production pipeline, an email intelligence platform, and an agent skill ledger, are documented on the systems index.
The method behind those systems: framing, tradeoffs, and follow-through, developed in full in the handbook.
Start with the system context, constraints, users, failure modes, and business pressure before choosing a technical direction.
Evidence FirstTurn ambiguous requirements into clear boundaries, contracts, tradeoffs, and implementation paths that teams can reason about.
Maintainable DesignShip the work, document the decisions, review the implementation, and make sure the operational reality matches the design.
Production MindsetThe same method, structured for teaching. Ten chapters, from the first ReAct loop to enforceable safety guardrails in production.
Current thinking in long form. Case studies and frameworks from the systems above, written while the work is still in motion.
AI Agent Enabler means enabling the reliable adoption and integration of AI agents into products, workflows, and engineering practices, making them dependable parts of real systems rather than demos. Inside a company, this work maps to senior and staff backend, platform, or distributed-systems engineering roles.
I am a Senior Software Engineer with more than 15 years of experience building backend systems and modernizing complex architectures. My work focuses on reducing coupling, establishing clear service boundaries, and improving the scalability and reliability of platforms.
That work was built at Vinted, PlusDental, and Wayfair, spanning marketplace platforms, API migration, and supply-chain systems. The selected impact below names the scope; the full timeline follows.
Outside my professional roles, I build systems to explore these ideas in practice: ScaleShop, an interactive scalability lab; Tamoz, a stream-oriented agent project; and an AI-assisted film-production pipeline that produced two short films. I write about distributed systems, event-driven architecture, reliability, and practical AI-agent infrastructure.
Led the extraction of the platform's first independently operated microservice, backed by Kafka and a dedicated database, out of an existing monolith on a live platform. Work around it spanned parcel-tracking performance, carrier webhook integrations, and crawler infrastructure; influence extended beyond the service itself into engineering practices and technical hiring.
Tech Lead for a nine-month database-decoupling project; built the MVP for a decoupled Kafka consumer.
Zero-downtime API migration, internal platform tools, testing practices, and architecture improvements.
Questions about the public systems, the handbook, or a field note? Email is the direct route.