Estimated reading time: 5 minutes

Is there a chance I might do conference presentations in the foreseeable future? Maybe. Maybe not. Who’s to say!?

I thought of it during KubeCon. Can I collaborate with friends, explore opportunities with co-workers, and gain context from mentees, to improve my conference presentation chances?

I was at KubeCon Atlanta a few weeks ago. There was a very business-y presentation by consultants that felt more bullshit than a technical presentation deserving the slot in KubeCon. The fact that they – the presenters – were able to up their profile by just bullshitting their way using ChatGPT, and that such people can get such easy entrance, gave me great excitement around the possibilities of me presenting at similar conferences. The bar is set so low it seems, that the work we’re doing at my org is actually quite cutting-edge!

I shared the realization that we’re quite well-placed for a high-profile conference presentation with my team. Since our workstream has been so diverse, we could potentially talk on a diverse range of topics. So many potential conferences! I reached out to Mz – a mentee – and encouraged her to record her thoughts, ideas, plans, and experiences. And now she has an idea for a potential conference presentations.

How realistic is it that I – or my team – might be able to give a conference presentation in the next year. What’s the likelihood, what are the blockers and how might I overcome the blockers?

The blockers, let’s start there. My company needs presentations to be run through their public relations team. Anything I present with my friends or not has to go through the company. Any collaboration with folks outside our company, regardless of how interesting a concept I want to present on is, is likely to be rejected because of enterprise rules.That’s a big woof. It’s not literally impossible, however, if there’s genuine synergy of ideas in the presentation.

I have talked to a friend in Business Consulting to explore work he has done with his clients he might be allowed to share. He’d share his consultant perspective, and I, my industry angle. I have also reached out to someone on a potential public-private partnership presentation around what the city is doing. Unfortunately, it turns out that it is not his particular department and not in his purview.

There is a lot of cool stuff happening within my company that I could talk about. We have the CloudForge Marketplace, and we have a fantastic pattern of teams maintaining inner source projects. We have the CFA project, which is pretty massive, and we could share it with enterprisey companies who’d find it very fascinating because it’s working at a massive scale. Then there is the Distilling concept that the squad I lead have been looking at. Then there was the Helm, Ansible, and CloudFormation deployment pattern that we were using a couple of years ago. We should have really shared that pattern somewhere. It was a crappy pattern in the end, but people have loved hearing the story of our infrastructure journey.

What are the other opportunities?

Everyone’s trying to optimize RAGs. If our squad can find a neat way to do a full-text search on the terabytes of data we have, people would be very, very excited. Simon Wilson suggests that instead of doing RAGs and chunking, and so forth, it is not a bad idea to let the agent planner generate the queries and run them. That saves embedding costs as well as storage costs. It saves us the work of the entire embedding pipeline, actually, and you’d only need full-text. You’d have the full-text search, plus semantic search for free thanks to the LLM layer in between . That would be something so exciting. That should be my next project. [ 10/16/2026 Edit: This did end up becoming the next exploration with a Data Science squad I led and managed, and let’s just say…I was overly optimistic in my assessment on how useful the approach might have been. There’s a reason everybody’s not doing this already, turns out… ]

Searching terabytes of textual data without chunking using agentic RAG—RAG in quotes. That pattern could generalize into multiple other projects.

My Data Science squad has been working on Distilling embedding models for CFA, but it’s not worked out very well. If we can get distillation going with a good validation set, that would be amazing. But we don’t have a clear understanding of our data, and the validation set coming from the other side is not entirely comprehensive, so we’d probably spend a long time in just figuring out the data. Maybe this is going to end up being more effort than it’s worth. The agentic RAG part seems really exciting, I must pursue it. [10/16/2026 Edit: I did. Didn’t get anywhere exciting. Ugh]

What else? Maybe one way of approaching this is to make a list of all the potential conferences—dozens and dozens of them. I could mark out which ones are interesting, and start from there. That sounds interesting. Instead of thinking about what I have to offer, I can start with understanding what they want and fulfill that need.

[10/16/2026 Edit: I’m still looking for conference presentation opportunities.]