The success of a company's data projects depends on the foresight of its management. Technology, organization and people have to be fine-tuned again and again so that data teams can function efficiently and develop a competitive advantage for their company. Let me advise you on this.
Example projectsI offer the planning and implementation of data science applications. The term "data science" is more of a collective word than a well-defined term. I'm a trained statistician and a seasoned expert in certain areas of machine learning. I have significant experience in classic applications (prediction, categorization, cluster analysis), Monte Carlo simulations and automated language processing (natural language processing). I have carried out numerous projects in these areas. I have solid theoretical background knowledge and am familiar with common software packages (Python or R).
When I talk about data warehousing, I refer to the consolidation of structured data from internal and external sources in a central database. I have converted, expanded or migrated numerous analytics databases internally and as part of my project work. My focus is on data warehousing for digital marketing and customer relation management. I plenty of experience working with Redshift, BigQuery, Snowflake and Postgres and know the advantages and disadvantages of the respective products very well.
I have many years of experience in the development and deployment of microservices (REST API) on Amazon Web Services and the Google Cloud Platform. Depending on the application, I can design and implement a data transformation or data science application as Docker-based web service or as serverless application. In that regard, I have broader competencies than a mere analyst or statistician. I can deploy data-driven applications which are highly available, cost-efficient and fail-safe. These skills allow me to work on equal terms with DevOps and developer teams.
The term "Data Science" is a collective term for data analysis that goes beyond reporting and requires programming skills. I am a trained statistician and a semi-skilled expert in certain areas of Machine Learning. I have a lot of experience in classical applications (prediction, categorization, cluster analysis) and automated language processing (Natural Language Processing). In these areas I have done numerous projects, have solid theoretical background knowledge and am familiar with common software packages.
A company is growing and needs professional business intelligence to be able to work in a data-driven way. Reports in Excel or Google Sheets are to be replaced by dashboards in a BI solution. I advise on the selection of suitable technology and the establishment of an internal data team.
I have already carried out this type of business intelligence projects in very many variants both as a consultant and as a developer. In particular, I am very good at showing the risks and opportunities that will result from certain decisions in the medium term.
A company's data is not systematically prepared in the data warehouse and is difficult for non-technicians to access. Together with the data team, I establish processes for an agile further development of the data warehouse and a secure conversion and expansion of the data model.
While data integration is a purely technical problem, data modeling requires a high level of business and technical knowledge. Well-modeled data can be accessed by non-technical people via SQL or create their own reports in a BI solution.
A company wants to use data from the data warehouse or an external source in machine learning models. I advise on suitable technology to train the models and make them available as a microservice for integration into existing processes.
My technical focus is on classical statistics on tabular data (prediction, classification, clustering) and natural language processing (text classification, text generation). However, I can also include services for processing images and sound recordings in projects.
A company uses AWS or GCP and needs a service that does some form of data integration and requires the lowest possible cost of ownership. I advise on appropriate technology and offer implementation as a serverless application.
Serverless applications are particularly suitable for the development of services that are not permanently called. In combination with serverless databases, very cost-effective solutions can be created that are also worthwhile for small companies.
I am happy to offer a non-binding and free consultation. In this conversation we can discuss requirements, clarify questions and divide the project into smaller sections. You can then review your project in peace and decide whether you want to start with an implementation.
If you have the concrete intention to commission me with your project plan, I will create a backlog in table form for your project plan following our preliminary discussions. For the individual user stories I estimate the amount of work in time according to our current state of knowledge. The more circumstances we can clarify and define at the beginning of the planning, the more precise the project plan will be. Based on the backlog, the functional and technical work packages can be easily understood. When commissioned, I will import the backlog from the spreadsheet into a project management tool and continue it there together with you.
As a rule, I work on the basis of a fixed hourly rate rather than a flat rate in the sense of a contract for work and services. This means that I don't have to add any surcharges for unforeseen risks at the start of the project. I deliver a backlog with work packages and a transparent effort estimate. I show you the actual effort expended by the hour and will be happy to explain any significant deviations from the effort estimate. The project can thus be managed in an agile manner and in close consultation with you: We can make extensions or changes in the course of the project at any time. If parts of the project involve major implementation risks, I will of course point this out to you in advance and make more conservative effort estimates accordingly.
My clients benefit from the confidence and foresight I bring to their data projects through my experience and pay a competitive hourly rate for it. However, your project may be so simple to implement that you don't need this experience at all. If you can manage your project yourself, assess your risks and implement it internally, it would be a waste of money to buy in an additional external consultant with my skills.
Normally, I import the agreed backlog into a Jira instance, which you are welcome to access yourself. I present the project progress to you at fixed project management meetings based on the completed user stories. We discuss the next steps, set priorities and discuss dependencies. I also inform you about the current status of the project budget at each meeting. If you prefer another project management software, we are happy to use it, as long as it does not cause disproportionate additional work.
In many projects I deliver the desired solution directly into the IT infrastructure of my customers. I am an experienced and certified developer for Amazon Web Services and the Google Cloud Platform. In these environments I develop complete, serverless applications including infrastructure (as code) and going live. If customers use another cloud platform (e.g. Microsoft Azure) or their own data center, I can deploy the applications as Docker containers, but leave infrastructure and operation to my customers.
I am an experienced and certified developer for Amazon Web Services. For projects in AWS I offer continued support after going live. For this purpose, a monthly maintenance budget has to be agreed upon, which will only be used if needed. I provide hourly reports on the use of the maintenance budget and will gladly explain the scope and reason for the work performed. A maintenance budget can also be used to improve the documentation, testability, security and maintainability of an existing solution in order to reduce risks and costs in the long run or to prepare the project for a seamless handover to another data engineer. If I cannot take over the maintenance, I will of course inform you in advance and support you in finding another data engineer.
In my projects I almost exclusively use Python for backend programming and if needed Javascript/Typescript or Java/Scala. I always deliver my Python projects for Python 3.9 (or higher) with typing and testing. Packages I develop are syntax and type checked with flake8 and mypy. Tests I deliver with pytest, where the degree of test coverage depends on the specific project. For serverless applications on AWS, I use SAM and deliver my projects with a CI/CD pipeline so that the application can be easily developed and deployed by third parties. If the application is to be run elsewhere, I can of course also deploy it as a Docker container.
You can reduce your planned project costs by hiring a less experienced developer. However, it's not the developer's hourly rate that matters, but the service provided within the budget. This includes good consulting and tailoring the project proposal to the available budget. We have to structure your project in advance in such a way that we can achieve your technical goals and at the same time control your risks. This is often not easy and requires experience and foresight. When choosing your external service provider, keep in mind that not every developer is a consultant and not every consultant is a developer. I offer a combination of both roles, ensuring transparent project execution and clear communication with budget owners.