Super Data Science: ML & AI Podcast with Jon Krohn 986: Building Hardware is Hard but AI Agents Help, with Kishore Subramanian
18 snips
Apr 24, 2026 
Kishore Subramanian, CTO of Propel Software and former Google Assistant engineer, discusses building Propel One AI and enterprise agents on Salesforce Agent Force 360. He covers how PLM and QMS support hardware teams. He explains agentic automation for reviews, security tradeoffs, and practical steps to ship AI tools. He also shares how yoga and meditation sharpen creative leadership.
AI Snips
Chapters
Books
Transcript
Episode notes
What Product Lifecycle Management Actually Does
- PLM is the system of record that carries a physical product from concept through manufacturing to the customer.
- It stores bill of materials, part hierarchies, manufacturing instructions, and post-sale quality data for hardware products.
Extensible Data Model Lets One PLM Fit Many Industries
- Extensibility enables one PLM platform to serve high-tech, medtech, and consumer goods by letting companies model domain-specific data and processes.
- Propel builds on Salesforce so customers can customize data models, processes, and UX per vertical while meeting regulated medtech requirements.
Change Review Agents Catch Hardware Pull Request Risks
- Propel One AI combines product data, documents, and metadata into a Data360 lake and layers generative and agentic capabilities.
- Example: change-review agents inspect engineering change orders (hardware pull requests) to surface compliance and downstream risk early.
