As IT Consultant I am wearing many hats when embarking on a project. Due to the confidential nature of my assignments it is often not possible to talk about what I did, when, for which client. To give you an idea what I can provide let me introduce you to a set of anonymized case studies of real projects I successfully completed.
Organizational and Product
Challenge: Slow and inflexible IT department
Solution
- Due diligence and analysis together with client before any action.
- Reorganization of department into product teams.
- Emphasis on agile practices like Scrum, continuous delivery and lean product development, Pre- and Post-Mortems, Tech Radar.
- Case-study-based hiring process.
- Creation of common guidelines for the teams (Rest, QA).
- Outcome Much faster feature development. Less bugs and improved customer satisfaction. Happy teams.
Challenge: Constantly failing product release dates for B2C application
Solution
- Change of culture to allow teams to raise concern openly so that timelines can be adjusted. Honest prioritizing of features.
- Lean product development.
- Going live with a subset of features and developing features together with customer.
Outcome
Product roadmap that can be used by Sales and Market for planning purposes. Successful product release.
Challenge: Creation of B2B software product that failed before
Solution
- Design sprint with customers to test and verify ideas.
- Using lean practices to prioritize features.
- Moving team into full-stack direction using React and NodeJs.
- Questioning requirements and keeping things as simple as possible.
Outcome
Application went live with a stable team that maintains and implements features. The automated processes free-up resources and save time for both the company and the customers.
Tech
Challenge: High performance big data analysis pipeline in crisis mode
Solution
- Analysis of real problems and drafting a technical roadmap to go to market as soon as possible with the most important features first.
- Replacement of batched approach (Spark, Hadoop) with in-memory real time calculation and aggregation. Outcome was faster development pace and quicker bug fixes.
- Hands-on developed with a team of the client for long term maintenance.
- Usage of standard components (PostgreSQL, Kafka, Spring Boot) so that the system can be run on AWS, but also on bare metal hardware for cost savings or performance increases.
- Stack: Kafka, Spark, Spring Boot, Java, Scala, PostgreSQL, Gatling, AWS, Datadog, DynamoDb
Challenge: Graph-based tag-management system fails because of too much load
Solution
- Analysis of database system and isolation of performance problem.
- Prototype with PostgreSQL and LTree data structure to store and retrieve data and fix performance problems.
- Performance tests using gatling to prove performance of new system.
- Migration of data and successful rollout to customers.
Stack
MySQL, PostgreSQL, Java, Php, AWS, RDS, Ltree, Gatling, Spring Boot, Graph data structures, Rest Api
Challenge: Software prototype of company ended up in production but has severe stability issues
Solution
- Prototype was too successful and the clients of the customer love it, but the services and software written was not intended to be scaled.
- Deep analysis of business requirements and current implementation.
- Prioritization of problems with highest impact.
- Redesign of services and re-implementation of unstable systems using standard components.
- Migration of data from NoSQL to SQL.
- Migration from bare-metal to AWS.
Stack
ElasticSearch, Php, Java, Javascript, Spring Boot, Tikka, Lucene, RDS, AWS, PostgreSQL, React
