10/15/2026 updated
100 % available
Senior Data Platform Engineer — Kafka, Flink, Kubernetes, on-prem/air-gapped
Ankara, Turkey
Only remote
MSc CS, Linnaeus (SE) 2012 · MSc IT, UPC Barcelona 2011 · BSc EEE, Başkent 2008About me
Senior data platform engineer, 14+ years. Kafka, Flink, Dask, Kubernetes, PostgreSQL; on-prem and air-gapped systems (300-node Dask cluster, 100M records). Agentic delivery with eval, approval gates and cost control. Remote from Ankara, UTC+3, full EU overlap. Starts with a 2-day assessment.
AI AgentsCloud EngineeringPython (Programming Language)Software ArchitectureBackendKubernetesApache FlinkDaskApache KafkaStream AnalyticsDockerMicroservices
Distributed Data Engineering
Builds and operates large-scale distributed data systems using Dask, Apache Kafka, Kafka Streams, Apache Flink, Apache Druid, Spark, Elasticsearch, Redis, and PostgreSQL. Architected a 300-server Dask fingerprint identification system processing 100 million records, live in production.
Real-Time Streaming & Analytics
Design and implementation of real-time analytics stacks using Apache Flink, Kafka Streams, Apache Druid as serving layer, and Apache Superset. Experience includes GPS/sensor anomaly detection, configurable rule-based alerts, and high-throughput live traffic data hubs handling millions of messages.
Cloud & Platform Engineering
Deep experience with Kubernetes (on-premises and air-gapped) and Docker; hands-on experience with AWS (ECS Fargate, RDS, S3, CloudFront) and Google Cloud (Dataproc, BigQuery, Dataflow). Delivery of production workloads in high-security, air-gapped public-sector environments as well as cloud-native architectures.
AI Systems & Agentic Pipelines
Design and delivery of LLM-based applications and agentic pipelines with evaluation, approval gates, observability, and cost control. Built a reusable agentic development pipeline with 10 specialised agents, scoped permissions, commit guards, and CI on every PR.
Backend Development (Java, Python, Node.js)
Over 14 years of backend engineering with Java (8 years), Python, and Node.js: TCP/UDP listeners, message decode/encode pipelines, microservices, and stream analytics engines.
Engineering Leadership & Scrum
Team lead and Scrum Master across full product lifecycles. Technical proposals, effort models, client and stakeholder management, and coordination of cross-functional teams in defence-sector and scale-up environments.
Product Delivery (Web & iOS)
Architecture, specification, review, and release of Next.js/TypeScript and SwiftUI products through an agentic delivery pipeline; implementation handled by the pipeline rather than hands-on front-end coding.
CI/CD & DevOps
CI/CD pipelines, commit guards, automated quality gates, and infrastructure as code (Terraform) for deployments across cloud and air-gapped on-premises environments.
Experience
Professional experience10+ years
Freelancer since2026
Languages
EnglishFluentTurkishNative speaker
Project history
Product development, data platform architecture, agentic delivery.
- Architected and shipped a B2B marketplace (data model, search, SMS-verified onboarding, KVKK compliance); 27 architecture decisions on record; live in private beta.
- Architected and shipped a native iOS app (21 screens) on an edge API orchestrating image-generation and vision models.
- Created a reusable agentic development pipeline: 10 specialised agents with scoped permissions, commit guards, CI on every PR and two human gates. I own the architecture, written specifications, review and release; implementation runs on the pipeline.
Defence-sector prime contractor - high-volume systems for major public institutions, on-premises and fully air-gapped
- Led the engineering team and coordinated with peer teams to design, deploy and operate production systems under strict governmental requirements; Scrum Master across the full product lifecycle.
- Designed and delivered the cloud backend of a large-scale video analytics system deployed across major government clients; evolved it from a single-appliance stack into a master/worker cluster without breaking units already in the field, leading both design and implementation.
- Operated all production workloads on Kubernetes across air-gapped on-premises clusters in high-security environments.
- Architected and built a fingerprint identification system on a 300-server Dask distributed cluster: 100 million records, 8 matching algorithms, results in 2–30 seconds depending on match type. Live and operational.
- Contributed to architecture design sessions and deployment strategies for expanding coverage to additional government facilities.