INWT Statistics

INWT Statistics is a German data science consultancy that has been supporting clients with tailored data analysis and AI solutions since 2011. The company positions itself as a full-stack data science partner, covering the entire lifecycle of data projects — from data validation and analysis, through model development and quality assurance, to production deployment, MLOps operations and long-term support. Its stated mission is to advance businesses and people through data. The firm organises its offering into service packages that span data validation and analysis, reusable code, impact assessment, technology recommendations, production roadmaps, development, quality assurance, MLOps operations with SLAs, refactoring and technical debt reduction, scaling and modularisation, cloud migration, feature expansion, and cooperation or handover to client teams. This end-to-end approach is aimed at organisations that want to build or modernise data-driven products and platforms. INWT's competence portfolio reflects more than a decade of experience in the data science industry. It includes Bayesian statistics, foundation models such as TabPFN, gradient boosting, LLM fine-tuning with Torchtune and Transformers, work with LLMs from OpenAI, Llama, Mistral and Hugging Face, LLM serving with Ollama and vLLM, machine learning and explainable AI, mixed effects models, multivariate analysis, natural language processing, RAG with Langchain and Qdrant, regression and classification, supervised and unsupervised learning, survival and time-to-event analysis, APIs, automated reporting, and dashboards built with Shiny, Redash or Vue.js. On the data engineering and platform side, INWT covers data analytics, data warehouses and lakes, data visualisation, ETLs, monitoring, Python and R, relational databases, AWS, Azure and Hetzner, CI/CD, cloud services and SaaS, CUDA and GPU computing, data platform work, DevOps and MLOps, Docker infrastructure, Kubernetes, microservices, monitoring, serverless computing and SLAs. The company also publishes podcasts, blog posts and open source projects, and names Till Bethge and Jan Blechschmidt among its data scientists. The team describes itself as broadly positioned, highly specialised and passionate, and states that honesty, integrity and orientation towards the highest standards shape its work and client relationships. It also publicly commits to the values of democracy, tolerance and diversity, welcoming different perspectives and backgrounds as a source of better decisions. INWT is a privately held company and does not appear to be listed on a stock exchange.

INWT Statistics was founded in 2011 and has since grown into a German data science consultancy serving clients with tailored data analysis and AI solutions. From its beginnings, the company has focused on applying statistical and machine learning methods to real business problems, and over more than a decade it has built a broad competence portfolio spanning statistics, machine learning, data engineering and cloud operations. The company describes itself as a full-stack data science team, meaning it covers the entire path from data validation and analysis through model development and quality assurance to production deployment, MLOps operations and support. Its service packages include data validation and analysis, reusable code, impact assessment, technology recommendations, production roadmaps, development, quality assurance, MLOps operations with SLAs, refactoring and technical debt reduction, scaling and modularisation, cloud migration, feature expansion, and cooperation or handover to client teams. Over time, INWT's technical focus has expanded alongside the data science field. Its portfolio now includes Bayesian statistics, foundation models such as TabPFN, gradient boosting, LLM fine-tuning with Torchtune and Transformers, work with LLMs from OpenAI, Llama, Mistral and Hugging Face, LLM serving with Ollama and vLLM, machine learning and explainable AI, mixed effects models, multivariate analysis, natural language processing, RAG with Langchain and Qdrant, regression and classification, supervised and unsupervised learning, survival and time-to-event analysis, APIs, automated reporting, and dashboards built with Shiny, Redash or Vue.js. On the platform side, it covers data analytics, data warehouses and lakes, data visualisation, ETLs, monitoring, Python and R, relational databases, AWS, Azure and Hetzner, CI/CD, cloud services and SaaS, CUDA and GPU computing, data platform work, DevOps and MLOps, Docker infrastructure, Kubernetes, microservices, monitoring, serverless computing and SLAs. The company also engages in knowledge sharing through podcasts, blog posts and open source projects, and its team members include data scientists such as Till Bethge and Jan Blechschmidt, as well as data engineer Jan Abel. INWT states that honesty, integrity and orientation towards the highest standards shape its work and client relationships, and it publicly commits to the values of democracy, tolerance and diversity. The firm remains privately held and active.

• 2011

  INWT Statistics founded; begins supporting clients with tailored data analysis and AI solutions.

Full-stack data science consultancy for tailored, production-grade data and AI solutions.

• Data validation and analysis

  Service package covering validation and analysis of client data.

• Production roadmaps and MLOps operations

  Roadmaps to production plus MLOps operations support with SLAs.

• LLM fine-tuning and serving

  Fine-tuning with Torchtune and Transformers; serving with Ollama and vLLM.

• RAG solutions

  Retrieval-augmented generation using Langchain and Qdrant.

• Dashboards and automated reporting

  Dashboards built with Shiny, Redash or Vue.js, plus automated reporting.

1. 1INWT | Data Science Beratung | PoC, MVP und Legacyinwt-statistics.de
2. 2Full Stack Data Science Team | Expert*innen für Datenanalyse und KI-Lösungeninwt-statistics.de
3. 3INWT Statisticswikidata.org