Benjamin Pohl
About Me
“Every challenge fuels the journey towards innovation and excellence.”
Accomplished Lead Machine Learning Engineer with over 7 years of experience in architecting and managing analytical platforms across Azure, GCP, and AWS, with a primary focus on Azure. Specializing in designing and deploying scalable machine learning models, I excel in leveraging advanced analytics to drive business insights and decision-making. Passionate about generative AI and staying ahead of emerging technologies, I am dedicated to creating innovative solutions that align with privacy and responsible AI development. My expertise ensures that projects are cutting-edge and future-proofed for long-term success.
Experience
Mercedes-Benz AG (2023-Present) - Stuttgart, Germany
Lead Machine Learning Engineer, Research & Development
- Directing the development of innovative LLM solutions to enhance data interaction and code access. Lead Developer for a holistic vehicle data analytics project, optimizing workflows and steering a substantial cost-saving analysis.
- Enhanced Collaboration: Launched Talk2Repo, facilitating seamless communication between developers and inner source repositories at Mercedes. Utilized llamaindex, RAG Agent, Milvus, a Streamlit frontend and continue.dev plugin to provide developers with easy access to sample code and resources, significantly enhancing productivity and collaboration.
- Innovation: Founded Data.Insight.Chat, a pioneering multi-agent framework (AutoGen) that enables intuitive data interaction via chat-like interfaces. Leveraged a React frontend, allowing for dynamic SQL queries, responses, and visualizations, transforming data exploration.
- Vehicle Data Analytics: Led a team of 8 developers with a budget of €1.2 million to develop a holistic vehicle data analytics platform. Managed the codebase, pull requests, and release cycles, ensuring project milestones were met on time.
- Cost Savings: Conducted a performance allocation analysis using Databricks, achieving annual savings of €500,000 through advanced analytical techniques and cost-efficiency improvements.
EnBW Energie Baden-Württemberg (2020-2023) - Stuttgart & Cologne, Germany
Machine Learning Platform Architect (MLOps) & Epic Owner
- Founded and managed the analytical machine learning platform, overseeing daily operations, continuous advancements, and stakeholder engagement.
- Initiated and managed the development of an Azure-based analytical machine learning stack, pivotal in enabling projects generating €250 million+ in annual revenue.
- Oversaw end-to-end provisioning of the analytical platform for data scientists, including data pipelines and batch/API deployments.
- Implemented robust CI/CD and Git workflows, automating testing and deployment to ensure superior quality in machine learning development.
- Led discussions on MLOps best practices, mentored new developers and data scientists, and managed the MLOps portfolio across the company.
EY (Ernst & Young) (2017-2020) - Stuttgart, Germany
Data Science Consultant
EY (Ernst & Young) (2014-2017) - Stuttgart, Germany
Part-time Consultant | Dual Course of Study (B.A)
Education
Queen Mary University of London (2018-2019)
MSc. in Big Data Science (Distinction, 1.0)
- Thesis: Content based music recommendation. Implementation using novel two stage unsupervised and supervised approach to circumvent the cold start problem.
- Core Modules: Big Data Processing, Machine Learning, Data Mining, Machine Learning for Visual Data Analytics, Deep Learning and Computer Vision, Data Analytics and Cloud Computing.
Corporate State University Stuttgart (2014-2017)
B.A. in Accounting, Taxation and Commercial Law (2.2)
- Thesis: Implementation of a smart contract transfer pricing model based on distributed ledger system (blockchain) for an international company.
Certificates & Awards
Scholarship
- One year data professional development scholarship (EUR 20,000).
Certificates
- DeepLearning AI MLOps specialization
- Hasso Plattner Institute 5-day design thinking specialization
Awards
- AWS DeepRacer Pro League Competitor
- Second place at the TDWI Data Science challenge (hosted by TDWI & Deloitte)
Languages
- German (M. T.)
- English (Fluent)
- Chinese (Beginner)
Programming Languages
- Python ❤️
- SQL
- Scala
- HTML
- CSS
- JS
- LaTeX
Cloud Services, Packages & Frameworks
- Hyperscaler (Azure, AWS, GCP) ❤️
- Databricks
- Spark
- Git
- Scikit-Learn
- PyTorch
- Pandas
- FastAPI
- Streamlit
- Docker
- Llamaindex
- Hugging Face Transformers
- LobeChat
- Ollama