Transforming Audio-Video Content Management: Unleashing the Power of OpenAI Whisper and GPT at Egnyte

Narendra Patel (~narendrapatel)


2

Votes

Description:

Improving document and Audio/Video(A/V) search, subtitles and summarizations with OpenAI Whisper and GPT At Egnyte

At Egnyte, we manage petabytes of audio-video content, with terabytes of new material uploaded daily, resulting in an immense dataset. To enhance our customers' experience with this vast dataset, we harnessed the combined power of OpenAI Whisper and GPT using Python, resulting in the creation of significant features.

These encompass a range of functionalities, some of which are released for preview, while others are in the development stage:

  • Enabling easy searchability of audio-video content for swift retrieval of relevant results from extensive content libraries.
  • Enhancing the video experience by utilizing A/V transcripts as subtitles.
  • Providing concise summaries of audio-video content for quick insights.
  • Facilitating interactive conversations with A/V content to expedite information retrieval.
  • Identifying similar content through vector similarity search.

Our presentation structure encompasses:

  • Introduction: Setting the context.
  • Requirements Overview: Understanding the needs that drove our innovation.
  • Solution Deep Dive: Delving into the intricacies of our approach.
  • Implementation Journey: Mapping the path from concept to reality.
  • Overcoming Roadblocks: Addressing challenges faced along the way.
  • Achieving Production Readiness: Preparing for deployment.
  • Scaling Strategies: Managing the system's growth.
  • Feedback Utilization: Incorporating user input for refinement.
  • Q&A Session: Addressing queries from the audience.
  • Acknowledgments: Recognizing contributors and supporters.

With Egnyte's commitment to delivering a remarkable user experience, our integration of OpenAI Whisper and GPT, alongside Python, has yielded transformative results.

Prerequisites:

Basic understanding of programming principles. We will keep the presentation as simple as possible.

Speaker Info:

I am Narendra, currently working as a Senior DevOps Engineer at Egnyte. I have close to ten years of experience in different roles such as Developer, DevOps Specialist, SRE, and RPA (Robotics Process Automation). Apart from my work at Egnyte, I have also been involved in some open-source projects, won hackathons, and shared my technical insights through blogs.

With a master's degree in Computer Applications, I'm a big enthusiast when it comes to technology. However, I'm not just focused on coding – I'm also passionate about making a positive difference. I contribute my free time with a non-profit organization with the goal of making the world a better place.

Among other things, I am a certified scuba diver and have completed the Everest Base Camp trek. When I'm not busy with work, you'll find me enjoying fun moments with my dog, trying out marathons or treks, or taking on new exciting challenges / adventure activities.

Section: Data Science, AI & ML
Type: Talks
Target Audience: Beginner
Last Updated: