Dr. Forooz Shahbazi Avarvand

AI & Data Science Consultant

• Advise organizations on applying AI, machine learning, and data-driven solutions to industrial and manufacturing challenges.
• Design and implement machine learning and signal-processing solutions for quality control, anomaly detection, and process optimization.
• Collaborate with cross-functional teams to translate operational problems into scalable AI solutions.
• Deliver freelance, contract, and interim consulting projects across industrial environments.

Selected Project:

• Improved reliability of production processes by designing and implementing signal processing and machine learning algorithms for automated quality control in textile sensor systems.
• Collaborated with engineering and production stakeholders to translate operational requirements into robust, scalable solutions.
• Optimized existing signal-processing methods for optical sensor systems and developed novel solutions for unresolved challenges under strict real-time and hardware constraints, enabling deployment directly on production machines.
• Identified future machine learning opportunities by evaluating alternative modeling approaches and defining data requirements to improve the precision of quality-control algorithms.

Module Lead & Lecturer at HSLU, Rotkreuz, Switzerland

• Module Lead and Lecturer in the MSc Program for MLOps at Lucerne University of Applied Sciences and Arts, Lucerne, Switzerland. Feb 2026–now
• Module Lead and Lecturer in the BSc Program for Data Engineering: March 2025–now
• Identified a critical gap in the curriculum and initiated the introduction of a new Data Engineering module, aligning academic training with industry demands.
• Designed and delivered the module from the ground up, equipping students with practical skills in data pipelines, scalable systems, and modern data infrastructure on the cloud.
• Module Lead & Lecturer in MSc for NLP: teaching transformers, semantic search, and LLM fine-tuning. 2024–now
• Established and directed an international module in partnership with academia and industry in Toronto, aligning real-world project design with educational goals and leading full program execution across technical, scientific, and operational domains.
• Supervised MSc and BSc theses in applied machine learning and NLP, including enabling LLMs to cite their results and provide transparent outputs, and developing and fine-tuning LLM solutions for legal applications.
• Guest lecturer and external expert for NLP at HSLU, Switzerland. May–September 2024
• Deputy head of AI and ML bachelor program at HSLU (ad interim). October 2024–August 2025, Feb 2026–July 2026

Senior Data Scientist at Bayer AG, Language AI, Berlin, Germany

• Designed and deployed an end-to-end AI-driven platform for insight generation, enabling stakeholders to monitor product trends, evaluate post-launch risks, and make data-informed decisions at scale.
• Led technical strategy for Generative AI initiatives by connecting data science teams and business stakeholders and translating business needs into technical requirements.
• Mentored junior NLP engineers, improving team productivity and accelerating delivery of high-quality models through knowledge sharing and technical guidance.
• Supervised a master thesis on fine-tuning large language models (LLMs) with in-house proprietary data at Bayer AG.
• Deployed production-ready REST APIs for NLP inference using cloud infrastructure.

Research Scientist at Alexa AI, Amazon, Berlin, Germany

• Collaborated with cross-functional teams to integrate NLP models at scale into Alexa, ensuring alignment with product requirements and seamless deployment.
• Built and optimized NLU models for multi-task dialogue understanding.
• Developed error analysis pipelines to identify misclassifications, enabling targeted model improvements and increased robustness in production systems.

Senior Data Scientist at Siemens R&D, Digital Factory and Mobility, Germany

• Reduced production costs and improved product quality by developing machine learning solutions for real-time detection of defects in sensor data, enabling automated quality control within manufacturing processes.
• Created clear visualizations and reports to present results to engineers and management, supporting data-driven decision-making.
• Patent holder: “Anomaly detection in track circuit devices using machine learning” — developed ML-based methods for automated detection of abnormal behavior in railway track circuits, reducing reliance on manual signal inspection.

Postdoctoral Researcher, Fraunhofer Heinrich-Hertz Institute (HHI), Berlin, Germany

• Conducted laboratory EEG measurements and managed large-scale acquisition of multichannel sensor data.
• Applied advanced signal processing and statistical modeling to study brain responses to 3D visual stimuli.
• Developed and adapted Python scripts (numpy, scipy, pandas, statsmodels) for data cleaning, feature extraction, and visualization of results.
• Documented and presented findings to cross-disciplinary research teams, strengthening skills in scientific communication and structured reporting.