How I Work

I am a solutions architect. My work sits where architecture meets operations: I design distributed systems, then make sure they can actually be run — measured, scaled, recovered and handed over.

I have worked across aviation, e-commerce, ride-hailing and telecom. The through-line is always the same: take a system that works in one team’s hands and make it work as a property of the whole organization.

For the past two years I have also been building the tooling my own discipline runs on, and that increasingly means agents — skills that turn raw engineering input into architecture decision records, review notes and performance analyses.

How I work

  1. Architecture is a decision log, not a diagram

    Every meaningful choice becomes an ADR with the alternatives and the consequences written down. Diagrams follow the decisions, and they are generated from source wherever possible so they cannot drift.

  2. Non-functional requirements need owners, not intentions

    A quality attribute that is not scored, gated and assigned to someone is a preference. I turn them into checklists with targets and verification methods, then into CI gates.

  3. Performance work is forensic, not heroic

    Baseline first, change one variable at a time, measure again. When total response time far exceeds downstream latency the problem is in an intermediate layer; when memory and CPU look fine but responses are slow it is almost always concurrency, not hardware.

  4. Technical success is not enough for adoption

    A shared library or platform service only lands if it ships with installation, documentation, support and operational guidance. I treat internal platforms as products with users.

  5. Teach the principle, not the implementation

    Educational material aimed at engineers works when it starts from the underlying concept, in a question-driven format, rather than from this quarter’s business case.

Where I have worked

  1. 07/2025 —

    Solutions Architect

    Turkish Technology — Turkish Airlines’ technology arm · Istanbul

    An NDC-based airline retailing platform is being built here — offer creation, pricing, basket, order — replacing legacy distribution with a modular, standards-aligned system. I joined to own architecture for the rule engine and promotion domains, and now lead NFR architecture, production readiness and the performance function across the whole estate.

    • Derived a company-wide NFR quality framework from ISO/IEC 25010: 11 categories broken into roughly 46 checklist items on a strict one-action-per-item principle, each with a target, a verification method and an owner.
    • Designed the go-live checklist and its scoring model, covering pipeline coverage per environment, IaC and Helm completeness, GitOps sync, secret management, canary and blue/green strategy, zero-downtime rolling updates, runbooks per failure scenario, incident process, SLA/SLO definitions and verified horizontal scale-out.
    • Drove readiness alignment across 100+ microservices owned by 15+ teams, defining for each quality attribute what evidence counts, who verifies it and how it is tracked — resilience, contract testing, scaling behaviour and SLO targets became per-service tracked properties rather than per-team judgement calls.
    • Automated the scoring: a reusable CI workflow evaluates each repository against centrally-owned matchers and emits a versioned readiness report; a central aggregator normalizes across schema versions, applies weights and renders a platform-wide scorecard.
    • Separated policy flowing down (gate rules, weights, blocking versus advisory intent) from results flowing up (the report contract), so org-wide standards change in one place rather than through a fan-out of pull requests.
    • Cut rule evaluation latency from 3.21 s to 30.6 ms P95 — a 105× improvement — by isolating the true bottlenecks rather than scaling hardware: stateless session pooling, JVM memory leak remediation, and thread-exhaustion analysis under contention.
    • Established a three-tier performance testing model (component, service, end-to-end) as an architecture decision, with a shared endpoint-level performance profile contract every service implements and execution environments provisioned as infrastructure rather than hand-built.
    • Decomposed the offer, pricing and basket domain into separate microservices, and designed a distributed rule evaluation architecture on Hazelcast that removed the single-node ceiling for a Drools engine carrying 30,000+ rules.
    • Established company-wide code review and architecture review processes, and authored ADRs covering Kafka error handling and retry/DLQ strategy, schema governance, observability alerting and database migration rollback.
    • Built the agentic toolchain the architecture function runs on: skills that generate MADR-format ADRs, RFCs, work items and architecture-level code reviews, plus a repo-based extraction pipeline regenerating C4 and Structurizr models directly from source with a generated/override file separation.

    Java · Spring Boot · Drools · Hazelcast · Kafka · PostgreSQL/CNPG · Couchbase · Keycloak · OpenTelemetry · Grafana · Prometheus · K6 · Kubernetes/OpenShift · Terraform · ArgoCD · Backstage · APISIX

  2. 08/2024 — 07/2025

    Software Development Engineer

    Trendyol Group · Istanbul

    Returned to Trendyol to lead architecture on the international search and content platform.

    • Owned services sustaining 2M req/min under campaign-peak traffic — holding consistency across service boundaries with the Outbox pattern so state changes and their events publish atomically, and guarding throughput with circuit breaking, rate limiting and retry.
    • Led the architecture for multi-language translation of 200M+ documents, reducing document count by 30% through index-time embedding rather than storing a document per language.
    • Implemented custom ArgoCD metrics driving intelligent auto-scaling, replacing static CPU thresholds with signals that reflect actual workload.
    • Decoupled services along domain-driven boundaries, reducing cycle time by 25%; delivered DDD and Event Storming workshops.

    GoLang · Kafka · ElasticSearch · Couchbase · ArgoCD · Kubernetes

  3. 09/2022 — 08/2024

    Software Development Engineer

    Heetch · Remote / Paris

    Ride-hailing platform operating across Europe and Africa; worked on the driver domain during its multi-country expansion.

    • Refactored legacy driver management into a multi-country scalable system, enabling expansion across African markets with country-specific rules isolated from the core lifecycle.
    • Introduced an API-First approach at team and then company level, with OpenAPI specifications and contract testing as the integration boundary.
    • Integrated Wave as a payment provider for West African markets, including multi-currency transaction flows.
    • Designed Avro-based event schemas for driver lifecycle events, and improved observability with Datadog SLOs and alerting.

    GoLang · gRPC · Avro · Protobuf · Kubernetes · AWS · Datadog

  4. 01/2021 — 09/2022

    Software Development Engineer

    Trendyol Group · Istanbul

    • Designed scalable data models in the Product domain serving millions of products under high-throughput read and write patterns.
    • Scaled Kafka event pipelines and ElasticSearch clusters carrying product search traffic.
    • Led CI/CD pipeline development, and mentored the QA team on microservice-compatible testing and contract testing.

    Spring Boot · Kafka · RabbitMQ · ElasticSearch · Couchbase · SQL Server

  5. 05/2019 — 01/2021

    Software Development Engineer

    Easy Software · Istanbul

    • Developed common microservices — gateway, configuration server, authentication — used across product teams.
    • Maintained the Kubernetes environment and Jenkins pipelines, providing CI/CD for the organization.
    • Delivered internal webinars on security, testing, refactoring and clean code.

    GoLang · Spring Cloud · Kubernetes · Jenkins · Keycloak · OAuth 2.0

  6. 07/2018 — 05/2019

    Software Development Engineer

    Adesso · Istanbul

    • Contributed to the architecture, reliability and scaling of a greenfield project.
    • Wrote design documents, led code reviews, and managed source control, builds and releases.

    Spring · Hibernate · ReactJS · MSSQL · Docker · Jenkins · Kubernetes

  7. 07/2016 — 07/2018

    Software Development Engineer

    Turkcell · Istanbul

    • Developed APIs for BiP, a large-scale distributed messaging platform.
    • Increased manageable message traffic by 10% by prioritizing third-party service calls.

    Spring · Kafka · Redis · Coherence · PL/SQL · Angular · Grafana · Prometheus

What I know

Architecture & design
Distributed systems, high-throughput transactional services, distributed consistency and idempotency, Domain-Driven Design, event-driven architecture, microservices patterns (Outbox, Saga, Circuit Breaker, Rate Limiting), Strangler Fig modernization, C4 / ArchiMate / Structurizr modelling, ADR and RFC governance
Production readiness
ISO/IEC 25010-based NFR frameworks, go-live checklists and scoring models, automated readiness gating in CI, SLA/SLO definition, on-call and incident management practice design, runbooks, release gates
Agentic engineering
Custom agent skills for engineering workflows, agentic test generation and analysis, automated architecture and documentation extraction from source, Spec-Driven Development, custom GPT development
Languages & frameworks
Java, Spring Boot, Spring Cloud, GoLang, Kotlin, Python, Node.js / NestJS, gRPC, ReactJS
Data & messaging
Kafka, RabbitMQ, ElasticSearch, Couchbase, PostgreSQL / CNPG, SQL Server, Redis, Hazelcast, Apicurio Schema Registry, Avro / Protobuf
Rules & decisioning
Drools, Rete-based evaluation, GoRules, distributed rule evaluation, stateless session strategies
Cloud & platform
AWS, Kubernetes / OpenShift, Docker, Terraform, ArgoCD, GitHub Actions, GitLab CI, Jenkins, Backstage, Spring Cloud Gateway, APISIX, Keycloak, OAuth 2.0
Observability
OpenTelemetry, Grafana, Prometheus, Datadog, K6, JVM performance forensics (thread stack analysis, GC tuning, connection pool optimization)
Practices
TDD, BDD, contract testing, API-First, Spec-Driven Development, DORA metrics, Event Storming, chaos and resilience testing
Domain
IATA NDC 24.1, ONE Order, offer and order management, airline retailing maturity, PSS / GDS / ATPCO integration

Education

  • M.S., Engineering Management (thesis stage)Yıldız Technical Universityexpected 2027
  • B.S., Computer Science EngineeringYıldız Technical University2015
  • TOGAF enterprise architecture trainingTurkish Technology, in-house programme

Get in touch

The fastest way to reach me is email. I read everything and answer what I can. I am always up for a conversation about architecture, production readiness and engineering with agents.