WORKSHOP
Mark My Words. And Pixels. A Hands-On Guide to AI Watermarking (WS14)
Date: 08 October 2026 | Time: 11:30 to 02:00 PM
Venue: Workshop Room 1, NIMHANS Convention Centre, Bangalore
FEES:
• Rs.299 for Leaders Pass holders
• Rs.1009 for Professionals Pass holders
• Rs.1249 for Knowledge Pass holders
• Rs.1699 for Community++ or Student Pass holders
• Rs.1999 for all others
(Limited seats available)
AI-generated content is everywhere. Let’s build something that leaves a mark, even when nobody can see it.
We’ll explore invisible watermarking using open-source models and libraries, build a working verification pipeline, and give it a hard time. Expect live demos, readable code, and a few watermarks that don’t survive the afternoon.
Mayank Raj
Staff Engineer, Infrastructure, Stripe
Shreya Agrahari
Analyst, Accenture
- Open-source contributors and maintainers, developers, AI engineers, architects, security practitioners, and curious students interested in adding watermarking to the software they build.
- Explore text watermarking with Google’s SynthID Text and Hugging Face Transformers.
- See image and video watermarking with Meta’s PixelSeal and VideoSeal, plus audio watermarking with AudioSeal.
- Connect frontier AI research with open-source implementations you can inspect and adapt.
- Understand watermarking, AI detection, and content provenance.
- Build an image watermarking and verification app using PyTorch and Gradio.
- Use Pillow and FFmpeg to explore how cropping, resizing, compression, and re-encoding affect detection.
Benefits/Takeaways of this workshop for the attendees (What will attendees do after attending the workshop which they were not able to do before attending this)
A working image watermarking app you can adapt for your own projects. You’ll learn to embed and detect watermarks, measure robustness, and understand what a detection result actually establishes.
Basic familiarity with Python and APIs. Bring a laptop with Python installed, permission to install packages, and internet access.
No prior watermarking experience needed. A healthy urge to break the demo is very welcome.
About Speakers
Mayank Raj is a Staff Engineer at Stripe whose experience spans distributed systems, hardware-backed security, data platforms and AI-assisted engineering. He focuses on the trade-offs that make complex systems reliable, secure and practical at scale. An OpenAI Codex Ambassador, he also runs SudoMeet, bringing this engineering depth to hands-on workshops and hackathons that help developers build with AI and open-source tools.
Shreya Agrahari works on making AI and agentic development accessible to a wider audience. Her focus is on helping people move from understanding what AI can do to building with it. Through public speaking on open-source tools, she brings a practical perspective to creating AI-powered applications and workflows