Backend Engineer @ Albert Heijn

Backend &
Distributed Systems

I build the identity & access layer for Albert Heijn's transport platform, a dedicated identity provider, policy-as-code, and per-app API gateways, and design the event-driven services around it.

Strongest where a messy domain problem turns into a clean, durable architectural decision.

Moving a microservices estate off a legacy monolith and one shared static-RBAC gateway onto policy-based, per-app access, while growing toward solution architecture.

View Work arrow_forward Download CV
tayam@core: ~
$ whoami

tayam-el-jebari

Backend & Distributed Systems Eng.

@ Albert Heijn · Transport Platform

Location: Haarlem, NL

$ stack --info
- Kotlin / Spring - Kafka / Debezium - Identity provider / OIDC - Policy-as-code authz - API gateway-as-code - Kubernetes
$ status
check_circle Building toward solution architecture.
$

Selected Work

All case studies arrow_forward
Case 01 Albert Heijn · current

Identity & access layer for a transport platform

Every service sat behind one shared static-RBAC gateway that tangled authentication with authorization. I turned the enterprise identity standard into a buildable transport-domain design and built the replacement: per-app, policy-based access with a dedicated identity provider, policy-as-code, and an API gateway provisioned as code.

Identity provider (OIDC/SAML)Policy-as-code authzAPI gateway-as-codeKotlin
Read case study arrow_forward
Case 02 Albert Heijn · Feb–Jun 2025

A reusable event-publishing framework

Every team was re-solving reliable event publishing on its own, some subtly wrong. I designed and built a reusable Kafka + Debezium framework that closes the dual-write hole: a projected ~16 of an estimated ~24 developer-years saved at full adoption.

KotlinSpring BootApache KafkaDebezium
Read case study arrow_forward
Case 03 Albert Heijn · recent

Using AI only where rules run out

One monolithic LLM prompt handled the entire classification step, brittle and hard to trust. I redesigned it into a routed service: deterministic policy where the answer is known, an LLM only where judgment is genuinely needed.

Multi-agent LLMKafka (event-driven)Classification & routingRegression testing
Read case study arrow_forward