North Africa / Multilingual products & operations

AI & full-stack engineer in North Africa.

Build across languages. Operate as one business.

Exceptional Software Engineer · Winner. North Africa.

AI & full-stack engineering / Oussama Zeddam

I’m Oussama Zeddam, an award-winning software engineer based in Algiers, North Africa. I help teams turn AI capability, business software and disconnected workflows into a product they can actually operate. From multilingual customer journeys to the APIs and permissions underneath them, I take senior ownership of the system as a whole.

What I build

Engineer for the complexity between markets.

Hire a senior full-stack developer for frontend, backend, APIs and AI integration. My work spans SaaS development, AI agent development, RAG systems and scoped fintech software engineering. Explore the Expensify fintech engagement.

Multilingual AI with clear boundaries

A useful assistant needs more than translated prompts. Its knowledge, retrieval, tool permissions and handoff paths must fit the business. I design the workflow around approved information and test the language and failure cases that matter to your users.

One platform. Coherent operations.

Bring customer-facing software, internal administration and integrations into a shared architecture. Model roles, data ownership and operational rules explicitly so adding a market or workflow does not mean rebuilding the product from scratch.

A technical partner across borders

Discuss requirements in Arabic, French or English. Agree on decision ownership, written milestones and communication windows before delivery starts. My base is Algeria; the working relationship is designed for regional and international collaboration.

Selected engagements and contributions

The work makes the case.

Official OpenAI Agents SDK launch artwork showing agent handoffs and guardrails on a blue abstract background

OpenAI Agents SDK / Open-source contribution

OpenAI — engineering the official Agents SDKs

Merged upstream · Python & JavaScript

Strengthening the infrastructure behind production AI agents: fail-fast validation against ambiguous tool dispatch, collision-resistant MCP tool names and application-controlled UI streaming. Implemented, regression-tested and merged into OpenAI’s official Python and JavaScript SDKs.

Read the case study
Stanford University production software case study on structured sentiment analysis

Stanford University / Selected production software delivery

Stanford University — AI for a global research leader

$2.3B in FY2025 sponsored research support, including SLAC

Selected production software delivery for Stanford University. Structured sentiment analysis, target extraction and evidence-grounded outputs, with abstention rules that keep unsupported inferences out of the decision path. The public technical note explains the architecture.

Kairaxis marketplace homepage showing intent-led discovery across AI skills, prompts and workflows

Kairaxis / Client production system

Kairaxis — a marketplace for the agent economy

MCP-ready · Public APIs · Client production platform

Architected and delivered a live AI marketplace for a client: searchable skills, prompts and workflows, inspectable provenance, permissions and compatibility, plus public APIs and authentication. A full-stack product built for people and agent clients.

AgentrixOS homepage showing an all-systems-online interface for autonomous business agents

AgentrixOS / Client production system

AgentrixOS — OpenAI, Anthropic & Google in one platform

Model integrations · Client production platform

Architected and delivered a client platform that brings multi-model AI agents into business operations. Orchestration, RAG, knowledge connectors, BYOK and guardrails form one system for deploying, supervising and integrating specialized agents.

How we work

Regional understanding. Evidence you can inspect.

  • Work directly with the engineer responsible for delivery
  • Agree on scope, milestones and acceptance criteria
  • Leave with working software and a usable handover

North African products often sit between languages, established business processes and international software providers. I start by mapping those boundaries: which system owns the data, which users can act, where information moves, and which decisions need human judgment.

I was voted Exceptional Software Engineer at The Connected Awards, North Africa edition, in 2022. The engineering record also includes public OpenAI Agents SDK contributions and selected production systems for international organizations. Those are distinct forms of evidence, with the scope of each engagement stated clearly.

For a regional launch, the first deliverable is a clear path to an operable release: product scope, integration constraints, language requirements, acceptance checks and a maintainable handover. We agree on availability for your team’s location rather than assuming every market works the same way.

Before the first conversation

Questions worth asking.

Where in North Africa are you based?

I am based in Algiers, Algeria. I collaborate remotely with regional and international teams; I do not present offices in other North African countries.

What award did you win in North Africa?

I was voted Exceptional Software Engineer at The Connected Awards, North Africa edition, in 2022. The organizer’s public winner record is linked in the recognition section on this page.

How do I find the best full-stack AI engineer in North Africa for my team?

Look for evidence across the complete delivery: product interfaces, backend systems, data access, AI behavior and production operations. Ask what the engineer personally built. My case studies describe the problem, my responsibility and the delivered system, with public sources where available.

Can we work together in French, Arabic or English?

Yes. I work in French, Arabic and English. A multilingual product still needs its own agreed language scope, representative evaluation cases and a clear escalation path; we define those during project discovery.

Your next production system

Bring the hard problem. Let’s build the way through.

Tell me what needs to work, what is blocking it and what success looks like. I’ll review the fit and help define the next move.

Discuss your project