By Adrian, who attended the event
Hello everyone! Adrian here, just back from the fantastic DevOps Oxford meetup we held last week. I’m excited to share all the highlights from the day, especially as we had a brilliant new venue, Inventa, kindly provided by Mission Street. A big thanks to them for hosting us!
Two Amazing Talks: DataOps and LLM API Calls
We were lucky enough to have two top-notch speakers on the agenda:
- Marco Zaccaria Di Fraia – DataOps in Action
- John James – Adding History to Your Prompt API Call
Each brought their own experiences and perspectives, resulting in a lively and inspiring session for everyone.
Talk 1: Code, Data, Kindness, and the Future – Marco Zaccaria Di Fraia
Marco kicked things off with his talk on applying DevOps principles to data ecosystems using a DataOps framework.
- Why DataOps Matters: With AI on the rise and digital transformation happening everywhere, organisations need reliable, fast-flowing data. Marco introduced DataOps practices that streamline data delivery, improve quality, and build trust in data-driven decisions.
- Emphasis on People: One of the biggest things I took away was Marco’s focus on the human side of DataOps. He explained how collaboration, empathy, and kindness create stronger bonds within teams, helping them solve problems more creatively.
- A Unique Perspective: Coming from a background in AI research, space exploration, and sustainable development, Marco offered fresh insights on how these methods apply across different industries.
If you’d like to dig deeper into DataOps principles, you might find the DataOps Manifesto a useful starting point. It’s amazing how a clear methodology can help your teams embrace continuous data integration and delivery. According to a recent Forrester study, adopting DataOps can cut data pipeline interruptions by up to 75% – food for thought if you’re struggling with slow or error-prone data workflows.
Talk 2: Adding History to Your Prompt API Call – John James
After Marco’s talk, John James took the stage to discuss a major challenge with large language models (LLMs) when using them via APIs: how to maintain chat history for context.
- The Issue: Unlike consumer-facing chat interfaces (e.g., ChatGPT itself), API calls can’t “remember” previous conversations. Developers have to manually pass chat history.
- Different Approaches: John outlined various ways to do this efficiently—each with its own cost, complexity, and performance trade-offs. He showed us a neat example from his work at Charisma.ai, where they use conversational AI for training and storytelling.
- Practical Tips: Everyone loved learning about how to store past messages without ballooning costs or slowing down performance. If you’re interested in best practices, definitely check out OpenAI’s ChatGPT API documentation to see how they suggest handling conversation context.
John’s mathematical and web development background shone through in the structured, data-driven approach he shared. It was a real eye-opener for any team that wants to integrate LLMs into their apps without losing the thread of the conversation.
Looking Ahead – Join Us in April!
We’re already preparing for our next DevOps Oxford event on 22nd April 2025, and I can’t wait to see what’s in store. If you fancy joining us, be sure to register here.
Want to Speak?
We’re always on the hunt for new speakers! If you’ve got experiences to share about DevOps, DataOps, AI, or any related field, we’d love to hear from you. Enquire here about speaking opportunities – let’s keep building this fantastic community together.
Looking for a New Role or Recruiting?
If you’re on the hunt for your next DevOps role or need help building your team, visit our homepage for open positions and expert recruitment support. We work with brilliant organisations and are here to help match talent with the right opportunities.
Stay Connected
Follow us on LinkedIn and keep an eye out for upcoming events – DevOps Oxford is all about sharing ideas, sparking conversations, and nurturing the tech community in our city. Thanks again to everyone who came along, and I hope to see you all in April!
#DevOps #DataOps #AI #Oxford #OxfordTech #TechEvents #HumandEvents #DevOpsEvent
