Stephen Vaillancourt

AI Product Strategist & Architect

• Drive mid- and long-term AI and product strategy, aligning technical vision with corporate objectives.
• Provide strategic guidance on AI integration and discrete engineering innovations across the product portfolio.
• Architect and implement a semantic layer for Windchill, enabling scalable, intelligent capabilities; developed a prototype text-to-SQL interface on top of this layer.
• Serve as SME for Codebeamer performance and scalability, leading workshops for 20+ participants, diagnosing systemic issues, and authoring a remediation plan adopted into the product backlog.
• Represented PTC on the NVIDIA Developer Council
• Actively track AI and data engineering trends, sharing insights to foster a culture of continuous learning and cross-functional innovation.

Innovation Architect & AI/Data Science Leader

• Partner with executives to identify future product innovation needs, researching and prototyping solutions using state-of-the-art data science and engineering tools.
• Factory Visual Analytics – Served as architect, developer, product manager, and data scientist for a system using chained ML models to label datasets and build an action recognition classifier, converting worker activities into analyzable time-series data for pattern and anomaly detection.
• Enterprise Data Platform – Chief Technical Architect & Evangelist for a multi-tenant data platform within PTC’s SaaS environment; defined architecture, championed business alignment, and became SME in modern data platform technologies over 18 months.
• Digital Thread Service – Designed a novel architecture for synchronizing related data across multiple PTC products; solution is now being implemented as the backbone for future digital thread offerings.
• Windchill Data Migration Tool – Rebuilt a large-scale SQL Server → Oracle migration tool from scratch in under two weeks using open-source components, eliminating ~$1M/year in licensing fees while increasing performance and functionality.
• CAD Similarity Search – Designed and developed a system to perform CAD file similarity searches via graph decomposition and graph neural networks; resulted in a filed patent (Use of Graph Neural Networks to Index Computer Aided Design Models Based on Geometric Characteristics, App. #18665538).
• Corporate Development – Supported M&A strategy by identifying acquisition targets and serving as technical SME during evaluations.
• Represented PTC on the NVIDIA Developer Council in multiple sessions.
• Performance Optimization – Collaborated with Frustum (PTC acquisition) to optimize CUDA-based generative design code, improving performance through C++, CUDA, and advanced profiling techniques.

Data Scientist – Service Operations Analytics

• Collaborated with business stakeholders to translate operational challenges into data-driven solutions, leveraging service operations data to deliver actionable insights and deploy predictive models.
• Led development of a paragraph2vec-based classification model predicting the most relevant knowledge articles for customer case resolution across five product lines.
• Achieved 30% case deflection rates (goal: 10%), delivering ~$10M/year in efficiency gains; recognized with a corporate Innovation Award.
• Models remained in production until 2025 due to sustained impact.

World wide backline for performance and stability problems of Windchill and enterprise Java application.

• Became subject matter expert in Oracle performance, root cause diagnostics across multiple technology tiers including a several million line codebase and intricacies of the Java runtime
• Developed tools and processes which became the backbone of the diagnostics performance and stability techniques still in use today, and which have been key in enabling the product's ARR growth to 1/2 billion dollars.
• Spread the knowledge on best practices and diagnostic techniques through creating and giving training classes, mentoring, coaching, onsite consulting, speaking at conferences and writing knowledge articles
• Performed performance tuning or root cause diagnostics on approximately twelve thousand different customer datasets

• Served as subject matter expert for two CAD data management database products, providing customer support for functional, runtime, and usability issues.
• Conducted troubleshooting, root cause analysis, onsite support, and training; developed training materials and mentored colleagues as a backline resource.
• Earned multiple top performance awards for exceptional customer service and technical expertise.