08/04/2026 updated

**** ******** ****
100 % available

STREAMING DATA ENGINEER · CONSULTANT · FLINK & KAFKA SPECIALIST

Negresti-Oas, Romania
Only remote
BACHELORS OF COMPUTER SCIENCE
Negresti-Oas, Romania
Only remote
BACHELORS OF COMPUTER SCIENCE

Profile attachments

Flaviu Cicio - CV.pdf

About me

Senior Data Engineer with 10 years of experience building near-real-time data pipelines and platforms. Expert in Apache Flink and Kafka, including CDC ingestion, schema management, stateful processing, and low-latency serving layers.

Amazon S3ElasticsearchProtocol BuffersPostgreSQLScalaSQL DatabasesData StreamingSnowflakeChange Data CaptureBackendGitlab-ciAWS AuroraDebeziumKubernetesApache FlinkAvroApache KafkaAWS EKSDocker
I’m a Senior Data Engineer with 10 years of experience and an Apache Flink Specialist focused on building production-grade real-time data streaming systems.

My background includes backend engineering, distributed systems, and data streaming architectures, working with Java, Apache Flink, Kafka, Spring, Kubernetes, and AWS.

Over the last five years, I’ve specialized in Flink, designing high-volume, low-latency pipelines, optimizing stateful processing, and helping teams build reliable streaming applications.

I work across architecture, implementation, performance optimization, deployment, and production support, while also providing technical guidance and Flink best practices.

Core areas:
• Apache Flink
• Apache Kafka
• Real-time data processing
• Streaming architecture
• Java backend systems
• Distributed systems

Experience

Professional experience5 - 10 years
Freelancer since2016

Languages

EnglishFluentRomanianNative speaker

Project history

Data Streaming Consultant (Apache Flink)

Discogs (online music database & marketplace)

Media & Publishing

Advisory engagement to unblock backfills of PyFlink CDC pipelines feeding Elasticsearch, where the main pipeline processed ~5 records/sec.
  • Cut full-backfill time from 10+ days to ~12 hours by redesigning the pipeline topology: reordered a 5-way join, eliminated per-record UDF parsing, and moved enrichment upstream, cutting state from multi-TB to ~1 TB.
  • Guided one of the first production migrations to Flink 2.2 with the ForSt disaggregated state backend (state in S3), making in-place rescaling routine and halving infrastructure cost.
  • Established correctness baselines (late data, replay idempotency), backfill runbooks, and reference deployment templates for future pipelines.
  • Technologies: Apache Flink 2.2, Python, PyFlink, ForSt, RocksDB, Apache Kafka, Debezium CDC, Elasticsearch, AWS EKS, S3, ArgoCD, Kubernetes

Data Engineer / Flink Specialist

Telecommunications provider, Netherlands

Telecommunications

The client needed near real-time order-status updates for customers, call-center agents, and internal systems through a centralized data hub.
  • Designed the streaming architecture that unifies order events from CRM and order-management systems into one intuitive customer-facing status.
  • Implemented Apache Flink jobs (Scala) that bootstrap state from historical data in Snowflake via Flink's HybridSource, then switch to the live Kafka stream.
  • Guided a team of 5 engineers in Flink and data-streaming best practices, code reviews, documentation, and handover.
  • Technologies: Scala, Apache Flink, Apache Kafka (AWS MSK), Avro + Schema Registry, Snowflake, AWS EMR, GitLab CI/CD

Data Engineer / Flink Specialist

Bank, Switzerland

Banking & Financial Services

The client was struggling to build stream-processing pipelines for a new payments and reporting platform.
  • Enabled data scientists to ship Flink SQL pipelines without writing Java by building a Flink SQL runner that parses scripts, injects configuration, and deploys jobs automatically via CI/CD.
  • Built a testing framework for Flink SQL (unit tests with mocked connectors via Apache Paimon; integration tests with SQL-based assertions against containerized Kafka), increasing development efficiency by 150%.
  • Supported 3 teams with production pipelines: Flink SQL internals, custom sinks, PyFlink jobs, plus training via workshops and pairing.
  • Speaker at Flink Forward Berlin 2024: "Testing Flink SQL Scripts: Simplifying Development for Non-Developers".
  • Technologies: Java 17, Python, PyFlink, Flink SQL, Apache Kafka, Kafka Streams, Apache Paimon, Avro, Protobuf, Apicurio Schema Registry, Kubernetes, Docker

Data Engineer, AI/Streaming

Stream-processing platform vendor

Internet & IT

The client wanted a proof-of-concept for running RAG (retrieval-augmented generation) pipelines natively on Apache Flink.
  • Built a real-time RAG data pipeline on Flink: Kafka-driven ingestion into a Qdrant vector store, retrieval, and generation stages, with chat memory kept in Flink state.
  • Implemented custom async lookup table connectors for Flink SQL (LLM and embedding), enabling pure-SQL RAG pipelines over Ollama, OpenAI, Gemini, and Amazon Titan via Langchain4j, later open-sourced.
  • Deployed and validated the pipelines on the client's cloud platform.
  • Technologies: Java 17, Apache Flink (DataStream + Table/SQL API), Flink SQL, Langchain4j, Qdrant, Apache Kafka, Docker

Data Engineer / Flink Engineer

Digital bank

Banking & Financial Services

The client needed CDC pipelines with schema governance and near real-time compliance reporting on its data platform.
  • Added JSON Schema support to the client's Flink SQL platform: a generator producing JSON Schemas from SQL DDLs plus a custom serialization format.
  • Delivered near real-time compliance reports in Flink SQL (new customers, high-risk customers, politically exposed persons) from requirements to production.
  • Implemented integration tests for the Flink CDC and Kafka-to-JDBC pipelines.
  • Technologies: Apache Flink, Flink CDC, Flink SQL, Apache Kafka, JSON Schema, JDBC

Senior Data Engineer

RWS Language Cloud, world-leading language services provider

Internet & IT

RWS's translation management system lacked self-service reporting; I worked in the 6-person Data Platform team that built it.
  • Built the near real-time reporting pipeline end to end: CDC from MongoDB into Kafka, Flink aggregations into PostgreSQL, served through Spring Boot services with an OData query API, for 20k+ clients.
  • Sustained peaks of 500k+ events per second; improved peak performance by 80% via state reduction, aggregation reordering, and state TTL.
  • Technologies: Java 11, Apache Flink, Apache Kafka, CDC, Spring Boot, OData, PostgreSQL, MongoDB, AWS (S3, Aurora, QuickSight)

Senior Software Engineer

Glofox, all-in-one software for fitness businesses

Internet & IT

Product engineering across booking, payments, membership, and onboarding flows (about 70% Java, 30% React).
  • Increased platform scalability and responsiveness by at least 50% by moving booking, payments, membership, and onboarding flows to an event-driven architecture on Kafka and SQS/SNS.
  • Accelerated delivery of new requirements by applying domain-driven design.
  • Technologies: Java 11, Apache Kafka, AWS (SQS, SNS, S3, Aurora, EC2), Spring Boot, PostgreSQL, MongoDB, React.js

Senior Backend Engineer

SecurEdge, Network-as-a-Service platform, US

Internet & IT

Remote contract for a US-based Network-as-a-Service company, in a team of 3 backend engineers.
  • Built the telemetry ingestion and test-scheduling services collecting 50+ TB of Wi-Fi metrics from 150 devices into MongoDB for 30+ US clients (universities, hospitals, and other campuses).
  • Made distributed network tests reliable and consistent with Python tooling.
  • Technologies: Java 11, Python, MongoDB, MySQL, Spring Boot, Hibernate, Morphia, Docker, Linux

Senior Backend Engineer

DaiVB, automotive, connected-car services

Automotive & Vehicle Manufacturing

Connected-car services for an automotive vehicle backend, migrating a legacy monolith to event-driven microservices.
  • Reduced vehicle-command latency from 3s to 500ms by moving communication from HTTP to MQTT, as part of migrating the monolith to ~25 event-driven microservices on RabbitMQ.
  • Cut infrastructure costs by 50% while improving scalability and availability.
  • Technologies: Java 8, Spring Cloud, RabbitMQ, MQTT, PostgreSQL, Hibernate, Jenkins

Contact form

Log in to get in touch

You need to be logged in to use the contact form.

Sign upLog in