Greater Hamburg Area
Hi, I'm Darius Murawski, Data Scientist who loves digging on the data challenges and build end-to-end solutions for them. This starts by understanding the business requirements, building prototypes in tools like jupyer notebooks and finish to the automated deployment of machine learning models on the cloud. I write about data science, data engineering and software engineering in general on my personal blog to share my knowledge and speak at meetups about it. I use python with pandas, sklearn, plotly, flask, jupyter, falcon, pytest, pelican, airflow with following tools: aws, jenkins, git, hdfs and Apache Spark with Scala
Promoted internally to serve as de-facto Solution Architect across two product teams — without formal line management. I own architectural governance, technology strategy, and AI enablement at department level. Key contributions: → Designed integration models and defined system boundaries across cloud, data platform, and AI layers for two cross-functional squads. → Removed organisational blockers through negotiation of access rights and cross-departmental dependencies — cutting bottlenecks that previously delayed delivery by weeks. → Enabled autonomous continuous deployment by guiding teams through internal technical certification, decoupling releases from rigid corporate release cycles. → Translated customer feedback into structured backlog items; coordinated requirement workshops and maintained shared product alignment across business and IT stakeholders. → Contributed to hiring: recent referral hire became a high-performer. Designed and delivered hands-on sessions on Architecture, Cloud, and LLM topics. Piloted GitHub Copilot group-wide as internal AI adoption multiplier. → Currently leading an on-prem to cloud migration of all productive use cases from OpenShift to AWS — driven by end-of-support timelines. New use cases already cloud-native; full migration of 3 productive systems planned by July 2026.
Embedded engineer in a 5-person cross-functional squad responsible for data products in the Motor insurance department. Dual mandate: rapid prototyping and technical delivery alongside strategic AI product evaluation. Key contributions: → Reduced pricing workflow cycle time by 60× (3 months → 3 days) by building a self-service automation pipeline — enabling non-technical business users to run analyses independently. → Cut LLM evaluation pipeline runtime by 20× (5 hours → 15 minutes) through smart caching, enabling rapid prompt iteration without burning API budget. → Designed a concept to replace high-volume call workflows with AI, projecting significant cost reduction at scale. → International stakeholder coordination with Generali Head Office; built REST API integration bridging legacy on-prem systems with cloud-native services. Stack: Python · Snowflake · AWS (S3, Glue) · OpenShift · Terraform · Jenkins / GitLab CI
Platform engineering role in a 9-person squad serving internal data science users. Delivered a zero-downtime re-architecture of the core data platform while handling live customer traffic. Key contributions: → Reduced release cycle from 2 hours to 20 minutes via CI/CD pipeline redesign — unblocking teams and increasing deployment confidence across the department. → Self-initiated and independently built an automated data governance tool: identified manual, error-prone checks as a day-one scaling bottleneck, proposed and delivered the solution without a formal mandate. Now live in production across teams. → Zero-downtime migration of 100% of customer data to new NFS storage and a new Ubuntu-based environment — replacing Red Hat Linux with multi-stage Docker builds on GitLab CI. → Enabled cloud connectivity to AWS: created firewall rules, routing logic, and onboarded AWS and Snowflake integrations into the on-prem platform. → Introduced data-driven retrospectives that measurably improved team process quality. Stack: Docker · OpenShift 4 · Python · Snowflake · AWS · GitLab CI
Volunteered as a Python teacher at ReDI School, an NGO supporting refugee and immigrant professionals seeking to enter the tech industry. Taught a comprehensive curriculum covering object-oriented programming (OOP), Docker, NumPy, testing methodologies, and additional software engineering fundamentals — equipping participants with practical, job-ready skills. The programme was honoured at the Bürgerfest Berlin, recognising the social impact of digital education for underrepresented communities.
Volunteered as a gymnastics coach for preschool children, organising and setting up obstacle courses and movement circuits adapted to young children's developmental needs. Responsible for planning each session, building up the parcour before training, and running activities in a safe, engaging environment with parents present. Strengthened skills in child-appropriate instruction, organisational reliability, and community engagement.
Full-stack data science role at Germany's largest independent data provider — covering everything from ETL architecture to GPU-accelerated model training, real-time inference APIs, and cross-department product leadership. Key contributions: → Shipped a production ML inference API handling 3,500 req/s on AWS — unlocking a new revenue-generating real-time targeting capability on previously untapped no-data traffic. → Migrated the batch scheduler from BICsuite to Apache Airflow — eliminating a significant five-figure annual enterprise license and retiring an end-of-life server, while improving SLA from <80% to 95%+ at zero additional tooling cost. → Automated GPU-accelerated model training (Gensim word2vec) via AWS EC2, cutting training cycles from days to hours. → Speaker at a Data Science Meetup and Fresenius University — a talk on real-time targeting architecture directly convinced a senior engineer to join the company, turning technical credibility into team growth. → End-to-end model ownership: trained linear models, ran hyperparameter experiments, and monitored production behaviour — actively retuning when data drift occurred due to a 60-day rolling deletion window on training data. Stack: Python · Apache Spark · Airflow · AWS · Cassandra · Gensim (GPU) · Grafana · Kibana
Senior engineering role with a dual mandate: continued full-stack web development on a large Ruby on Rails platform, and ownership of the company's first machine learning integration in production. Key contributions: → Led deployment of the company's first ML model into production — accountable for full integration into the Ruby on Rails monolith and sign-off for production launch. → Acted as interim Product Owner for the data quality team, including international travel to Lithuania for cross-border requirements engineering and solution evaluation with a distributed development team. Stack: Ruby on Rails · Elasticsearch · Python · RabbitMQ · Google Cloud
Full-stack web development on one of Germany's leading B2B marketplaces, building and maintaining high-traffic customer-facing services on a Ruby on Rails platform. Key contribution: → Owned the go-live of a monolith-to-microservice migration: managed URL remapping, ensured zero broken links, and achieved complete Google reindexing within 3 months with no measurable SEO impact — a technically and commercially critical delivery. Stack: Ruby on Rails · Jenkins · Microservices · Elasticsearch