Feras Allaou

I build developer tooling and AI-backed products — mostly TypeScript, NestJS, and infrastructure that deploys itself. 15+ years across backend, frontend, and cloud.

If you're in Lisbon, let's grab a cup of coffee. ☕

Deploy a Node.js app to AWS or Azure from a single config file — no hand-written IaC. HTTP apps, cron jobs, and queue workers, on Express and NestJS.

flowchart LR
    A["your app<br/>+ laranja.config.ts"] --> B["CLI<br/>compiles to IR"]
    B --> C["backend<br/>synthesizes"]
    C --> D["CloudFormation<br/>· AWS"]
    C --> E["Bicep / ARM<br/>· Azure"]
    D --> F["deployed"]
    E --> F

    style A fill:#FFDF7E,stroke:#1F1FE0,stroke-width:2px,color:#1F1FE0
    style B fill:#FDFBF2,stroke:#1F1FE0,stroke-width:2px,color:#1F1FE0
    style C fill:#FDFBF2,stroke:#1F1FE0,stroke-width:2px,color:#1F1FE0
    style D fill:#FFF,stroke:#FF9900,stroke-width:2px,color:#333
    style E fill:#FFF,stroke:#0078D4,stroke-width:2px,color:#333
    style F fill:#1F1FE0,stroke:#1F1FE0,stroke-width:2px,color:#FFF

The CLI never ships your cloud credentials anywhere — it compiles your app to a provider-neutral IR, the backend synthesizes the template, and your machine executes the deploy.

Plain English in, validated SQL out, executed against your live database. Every query runs in a Docker-isolated, tenant-scoped container with zero data retention.

A hands-on series on agentic AI, in Arabic — starting from how an LLM actually works, then tool chaining, memory, and getting agents to do real work instead of demos, in both TypeScript and Python.

▶️ Episode 1 — كيف يعمل الـ LLM فعليًا؟ Completions API و Tool Chaining · new episodes on @FerasAllaou. Companion repo coming soon.

Data orchestration across Salesforce, CMS, and AWS with NestJS + GraphQL + CDK. BNPL and PSD2 banking integrations. GraphQL and WebSocket backends for real-time GPS apps. Serverless on Lambda since 2019.

_{🇸🇾 Arabic · 🇬🇧 English · 🇹🇷 Türkçe · 🇪🇸 Español · 🇵🇹 Português — mentored founders at Impact Hub and Habitat Derneği}