Towards Data Science

favourite RAG system. Very closely. Somewhere inside that meticulously architected pipeline, a language model is producing rich, high-dimensional hidden states. Those states get compressed down into a string of characters. That string gets re-encoded, by a different neural network, into a different high-dimensional space. That vector gets stored. Later, another vector gets compared against it.

latest wave of articles, opinion pieces, and forecasts from consultants and AI vendors, all confidently predicting the imminent takeover of human work by agentic AI, I recently found myself thinking about my first year at Imperial College London, back in 2009.

is straightforward. You load the weights, load the data, and wait for it to finish. For most models, that’s fine, and the only real cost is time. When the wait gets too long, the usual fix is to add GPUs. Each one trains on a different slice of the data in parallel, and the work finishes faster. Nothing about the model state changes; you’ve just thrown more hands at it. That covers the common case, where the problem is time. Large models add a second problem that more hands can’t solve: space.

to interact with your coding agents is very important. There is a large variety of options out there that you can try, and they all impact how effectively you’re able to interact with your agents. I’ve spent a lot of time testing out different platforms to orchestrate coding agents, and in this article, I’ll give my opinion on some different tools and how you can find the tool that works best for you. I think it’s worth noting that the optimal tool will vary from person to person.

Loop engineering for hierarchical retrieval: reading a long document by its table of contents – gif by author of Part II of Enterprise Document Intelligence, a series that builds an enterprise RAG system from four bricks: document parsing, question parsing, retrieval, and generation. The retrieval brick is a trilogy: This companion (7quater) handles the case the brick meets in production: a document too long to read, with a table of contents too long to dump.

up a call to take a dinner reservation. A hopeful human voice: “I’d like a table for dinner at six,” the caller says. The AI agent answers without a pause: “Just to confirm, you would like to make a reservation for a dinner booking at 6 PM, correct?” The caller says yes, a table for four. “Understood, so that’s a party of four, also at 6 PM, is that right?” Yes.

some of the problems we face in implementing AI algorithms, we usually focus on the processors’ ability to handle them. But now, NVIDIA’s GPUs power the latest large language models, and companies compete to build faster AI accelerators. These new chips promise more computational power, more cores, and more operations per second. So, one would assume that the future of AI depends on building increasingly powerful processors! But there is a problem.

1. AI Everywhere. Workflows Untouched. on Agentic AI workflow enablement across enterprise teams, I kept seeing the same problem. Teams told me they were excited about AI. Then I watched their day-to-day: a workflow pulling from different data sources, multiple rounds of Excel copy-paste, and manual handoffs before anything reached a checkpoint. The most important intelligence rarely made it into any system.

between cholesterol and blood pressure is measured in 10 patients. The observed correlation is 0.62. Is this enough evidence that a relationship exists? Studies that measure 20,000 gene expression levels in 10 mice routinely apply multiple-testing corrections and stringent correlation thresholds. Researchers know that the more correlations they examine, the more likely they are to observe large values purely by chance, even when the genes are completely unrelated.

have a line like ESCALATION_THRESHOLD = 0.90. Above it, the agent acts. Below it, a human gets pinged. It’s simple. It’s tunable. It lets everyone feel like autonomy is under control. But it’s not the right solution. The fix isn’t a better number. The fix is noticing that the escalation threshold was never a percentage in the first place. It’s a price. Start with the wrong question Most teams ask a capability question. Can the agent write the SQL? Can it issue the refund? If yes, let it run.