Data-Driven Engineering: Production-Ready and Emerging Technologies

How to be Impactful with AI and ML Models in Practice

W​ednesday 25th & Thursday 26th February 2026

(3 hours 20 min each day)

07.00 (Los Angeles) | 10:00 (New York)
15:00 (London) | 16:00 (Berlin)

Join us for the 3rd annual ‘Exploring the Status of Data-Driven Engineering: Production-Ready and Emerging Technologies’ online seminar, where we'll focus on how AI and ML can have a real impact in practical product design.

Data-driven engineering methods are transforming how we design and improve products. Our seminar will highlight the methods that are useful and impactful in practice right now and those that are being developed. This is a key event for engineers, data scientists and machine learning practitioners keen to stay at the forefront of the technology, as well as for those who are just starting in the field. It is an opportunity to explore the cutting-edge technology and engage with the community shaping the future of data-driven engineering.

Why Attend?

• To gain real-world application insights. Acquire greater understanding of how data science and machine learning are effectively and impactfully applied to real-world problems. This will enable participants to apply these techniques to their own work and address practical challenges
• To stay informed about industry and academic trends. Discover the latest trends and technologies in engineering data science in industry and academia. This includes advancements in data science, machine learning, artificial intelligence and their real-world applications.
• To share knowledge and best practices and be inspired. Share the expertise and exchange best practices with fellow professionals to enhance understanding of production-ready and emerging data-driven technologies. The inspiration from the work of others will serve as a catalyst for achieving career/research goals and excellence in the field.

Agenda

Day 1 – Wednesday 25^th February

Production Ready Technologies

A​ll times below are in GMT (London)

15:00 Opening Day 1
Vladimir Balabanov, Boeing

15:05 Keynote Address: Engineering “Vision” with AI and In‑Silico: From Data to Design
Sonalee Tambat, Alcon

This talk covers how Alcon is embedding physics‑grounded AI to augment multiphysics simulation in the design loop to further compress ophthalmic device design cycles and breakdown engineering silos. The session highlights speed gains, infrastructure strategies and multi‑disciplinary collaboration.

15:50 Practical Considerations in Deployment of Machine Learning Models
Alex Lohse, Owens Corning Science & Technology

This talk explores real-world use cases where machine learning models, derived from engineering simulations, were successfully deployed within a manufacturing organization. We will discuss common pitfalls and lessons learned during adoption, how deployment strategies vary based on end-user needs, and how a single model can accelerate decision-making across multiple levels of the organization.

16:15 Short break

16:25 Structural Engineering Data Science for effective Production Operations
Carsten Buchholz, Rolls-Royce

While in the context of production processes a wealth of manufacturing and process data is generated structural engineering knowledge is often required to improve the process flows or find root causes in case of issues. The presentation will discuss options for using data science techniques in combination classical engineering simulation on the example of an aero engine build and balancing process.

16:50 GenAI for Engineering Design
Olcay Met & Fatma Kocer, Siemens DISW

In the last few years, we have started seeing value added applications of AI to engineering design. One of these applications is the use of geometric deep learning (GDL) for fast physics predictions. Depending on physics, GDL has accelerated performance predictions by 100-1000 times, enabling engineers to do design exploration and innovation which was not accessible with high fidelity simulations that take hours to days to complete.

In the forefront of AI for engineering design is now GenAI, which is a very challenging problem considering the fact that generated designs should not only be physically meaningful, but they should also be performing within constraints, and manufacturable.

In this presentation, we will be showing how GenAI for engineering design significantly accelerates product development and innovation. CarHoods 10k dataset is used for this purpose. This dataset is generated by Digital Design & Manufacturing Lab at The Ohio State University.

1​7:15 Short break

17:25 Workshop: Introductory ML: From Advanced Manufacturing Parameters to Fatigue Material Cards
Catherine Amodeo, Ford Motor Company and Alberto Ciampaglia, P​olitecnico di Torino

17:50 Discussion Session
Catherine Amodeo, Ford Motor Company and Alberto Ciampaglia, P​olitecnico di Torino

18:20 Close Day 1

Day 2 – Thursday 26^th February

E​merging Technologies

A​ll times below are in GMT (London)

15:00 Opening Day 2
Vladimir Balabanov, Boeing

15:05 Keynote Address: From Data to Design: Rethinking Engineering Design With Next-Gen AI
Faez Ahmed, MIT

Generative AI is transforming how we create, customize, and accelerate digital content. Yet applying these tools to engineering design introduces unique challenges, from maintaining precision under evolving requirements to working effectively in data-scarce environments and interpreting designer intent. In this talk, I will discuss these challenges and show how emerging engineering-focused foundation models are beginning to address them, reshaping workflows in areas such as vehicle design, CAD automation, and design optimization. I will highlight new opportunities enabled by generative AI that integrates multimodal data with engineering analysis and optimization, and present examples of AI-driven design co-pilots for complex engineering tasks. The talk will conclude with a forward-looking perspective on how AI can broaden design democratization, accelerate innovation cycles, and fundamentally reshape the role of engineers.

15:50 On Verification and Validation of Data-Driven Models for Time-Series
Saleh Rezaeiravesh, Manchester University

In line with broader developments in science and engineering, data-driven models for time-series prediction have received considerable attention in recent years. A wide range of approaches, from classical statistical models to modern deep-learning methods, are now routinely employed for forecasting time-dependent data. However, a standard and widely accepted framework for the verification and validation (V&V) of time-series predictions remains underdeveloped. This gap is particularly critical for physical and engineering systems, where the information content and temporal structure of the underlying processes must be faithfully preserved by predictive models. This talk reviews recent progress on V&V for time-series forecasting and highlights connections to uncertainty quantification and information-theoretic diagnostics. Furthermore, examples and analyses from turbulent flows will be presented.

16:15 Short break

16:25 Advancements in Multi-modal (image & text) LLMs to Answer Engineering and Technical Questions
Remi Duquette, Maya HTT Ltd

The presentation will focus on recent advancements in multimodal large language models and recent RAG algorithms that enhance the ability to leverage visual and textual information for improved “reasoning” to address complex engineering and technical questions. The results show how image-text fusion and engineering domain-adapted “reasoning” enable LLMs to interpret schematics, plots, and technical documentation, delivering context-aware, verifiable answers suitable for real-world engineering workflows.

16:50 Benchmarking Ontology-Guided Multimodal LLM Pipelines for Engineering Simulation Knowledge Curation
Olivia Fischer, Aerospace Systems Design Laboratory at Georgia Institute of Technology

This presentation examines the challenge of incomplete metadata in engineering simulation repositories and evaluates ontology-guided, multimodal LLM pipelines as a potential solution. The pipelines automate metadata extraction from both text and figures by combining OCR, vision–language models, vector embeddings, and retrieval-augmented generation. We explore how metadata tagging performance is affected by key design choices, including ontology depth, prompting strategies, and the selection of embedding models, vision models, and large language models. Results are evaluated by comparing automated annotations against human-curated metadata from AIAA- and IEEE-formatted simulation papers, using multi-label classification metrics to assess the impact of different pipeline configurations.

1​7:15 Short break

17:25 Optimizing Engineering Simulation with Centralized Data Management, AI, and the Digital Thread
John Williams , Rescale

The integration of High-Performance Computing (HPC), Artificial Intelligence (AI), and Computer-Aided Engineering (CAE) is revolutionizing engineering simulation. While these technologies offer immense opportunities, managing vast simulation data remains a key challenge. A data-centric approach is crucial for streamlining workflows, optimizing resources, and enabling AI-driven insights across industries like automotive, aerospace, and manufacturing.

At the heart of this approach is centralized data management—a unified platform that stores, organizes, and shares simulation data from various sources. This centralization addresses fragmented workflows and siloed data management, which often lead to inefficiencies and delays. By creating a centralized repository, engineers can collaborate more effectively, sharing insights and accelerating product development.

AI plays a pivotal role by automating data analysis and improving simulation accuracy. AI algorithms, including machine learning and deep learning, process large datasets, identify patterns, and generate predictions. In industries such as automotive, AI-powered simulations predict vehicle performance and safety, reducing development time and enhancing product quality.

To maximize these technologies, adopting a digital thread is essential. This continuous, traceable flow of simulation data across the entire lifecycle enables collaboration and iterative optimization. AI-driven workflows within the digital thread ensure simulations are constantly updated, improving efficiency and accuracy.

Cloud-based HPC solutions offer scalable resources, reducing the need for large upfront investments in infrastructure. These solutions allow companies to perform large-scale simulations cost-effectively without compromising on resource utilization.

Implementing AI-enhanced methodologies and centralized data management helps engineering teams overcome traditional CAE workflow barriers. Case studies from automotive, aerospace, and manufacturing industries show tangible benefits, such as reduced simulation runtimes, improved throughput, and faster decision-making. These advancements help organizations address complex design challenges and bring high-quality products to market faster.

17:50 Discussion session: Democratization of model generation and usage
Alex Lohse, Owens Corning Science & Technology

18:20 Close Day 2

This online seminar is being hosted by the NAFEMS Engineering Data Science Working Group.

U​niversity Student?
Contact jo.potts@nafems.org from your academic email address, for exclusive offers on registration for this event.

Sponsors

Support us ...

We would like to extend an invitation to your company to be part of this event. There are several outstanding opportunities available for your company to sponsor the seminar, giving you maximum exposure to a highly targeted audience of delegates, who are all directly involved in simulation, analysis, and design.

P​lease contact the event organiser
Jo Potts, for further information
jo.potts@nafems.org; +44 (0)1355 225688