Business Use Cases

Specialize in leveraging data science to drive impactful business transformations. Below are some key projects that successfully applied advanced analytics, machine learning, and deep learning (GenAI, LLM) to solve complex business challenges

Siemens s.r.o. [2024 - Present]

Senior Data Scientist – Siemens
Leadership & Advanced Analytics Expertise

Team Leadership: Managing and leading two high-performing teams within the Advanced Analytics Department:
- AI Framework Team: Developing an end-to-end AI framework to ensure safety, scalability, and robustness of AI solutions across the organization.
- AI Solutions Team: Delivering tailored AI solutions for diverse business lines, driving projects from problem definition to Proof of Concept (PoC), prototype development, and full-scale deployment.
Project Management: Preparing budget proposals, creating project timelines, and ensuring the successful, timely delivery of end-to-end AI solutions that meet customer needs.
Key Projects and Achievements:
- Advanced Search and Summarization with GenAI:
  Led a GenAI-based project for the Legal Department, creating a Retrieval-Augmented Generation (RAG) model that enables advanced search capabilities and summarization of complex information into layman’s terms.
- Factory Testing Process Optimization:
  Developed a solution to detect improper factory testing practices, resulting in €1.8M savings annually and reducing 3,000 hours of time wastage per year.
- Unusual Account Activity Tracking:
  Designed and implemented a global-level anomaly detection system to identify unusual activities, improving procurement processes and decision-making.
- Delivery Process Improvement:
  Created a solution to identify anomalies in sales order data across multiple countries. The solution, integrated via APIs into the customer's business ecosystem, enabled real-time anomaly detection and prevented false purchase orders.
AI Framework and Strategy:
Spearheading the development of a Global AI Framework to guide data science projects from inception to deployment. This includes fostering a robust strategy, framework, and operational processes (MLOps and LLMOps) for global data science teams to rapidly develop scalable and production-ready ML and GenAI solutions (PoC, MVP, and final product).
Team Impact:
Successfully matured the team's capabilities by implementing AI framework best practices, enabling the efficient development of robust AI solutions.

Zebra Technologies s.r.o. [2023-2024]

Data Science Leader - Zebra
Transforming Supply Chain and Repair Processes Through Innovative AI Solutions

Strategic Role: Reported directly to the Director of Data Science, driving the transformation of supply chain and repair processes using machine learning and deep learning strategies. Focused on developing Proofs of Concept (PoCs) and collaborating with cross-functional teams to implement impactful solutions, delivering strategic insights and boosting operational efficiency.
Steering Committee Leadership: Chaired a steering committee to introduce and standardize best practices in data science and foster a culture of innovation across teams.
Key Projects and Achievements:
- Identifying Alternative Parts Using NLP and LLM Models:
  Led the development of PoCs leveraging Natural Language Processing (NLP) and Large Language Models (LLM), including a Generative AI model to predict cost-effective alternative parts. This solution significantly reduced procurement costs and broadened product options.
- Enhanced Communication Efficiency with LLM:
  Designed and deployed a Generative AI tool to analyze supplier communications, prioritize critical threads, and improve response times by 40%. This strengthened supplier relationships and enhanced operational efficiency.
- Optimized Inventory Management with Data Science:
  - Phase 1: Developed a reusable Python library, the 'Safety Stock Determination Engine,' to calculate safety stock levels for suppliers and generate features for forecasting. This tool enabled data scientists to achieve consistent and reproducible results.
  - Phase 2: Created the 'Safety Stock Optimization Engine,' a Python library to optimize inventory policies, significantly reducing inventory costs for suppliers.
- Improved Repair Forecasting:
  Supervised a time-series forecasting initiative to predict global product repairs more accurately. This solution accelerated repair processes by 15% and minimized downtime, boosting operational efficiency.
- Order Lead Time Forecasting:
  Led a complex multiple-time-series forecasting project to predict delivery times for customer orders, factoring in exogenous features such as seasonal trends and logistical constraints. Leveraged MLflow for extensive model tuning and performance tracking, improving delivery reliability and optimizing inventory management.
- Attach Rate Calculation for the University of Texas:
  - Supervised five master’s students on a joint thesis project focused on calculating accessory attach rates and developing time-series models. Provided guidance on data science methodologies, Python, and Databricks, while mentoring students in effectively presenting their findings to non-technical stakeholders.
  - The project’s success led to its expansion into a full-fledged initiative, underscoring its business value and the practical relevance of the work.

Implemented Data Science Best Practices: Directed the adoption of robust data science practices, including comprehensive code reviews and MLOps. These best practices ensured high-quality outcomes and sustainable results across all data science projects.

Page updated

Google Sites

Report abuse