The world's leading publication for data science and artificial intelligence professionals. Submit an Article ✍️ contributor.insightmediagrou…
We're Global 🌏
- Tweets 64,384
- Following 1,916
- Followers 251,417
- Likes 7,954
When an agent fills the terminal with text, finding the information that actually matters can become a challenge. Eivind Kjosbakken explains how improving the way coding agents communicate can make workflows more effective.
towardsdatascience.com/how-t…
2
977
What happens when you reduce a dataset while keeping class separation in focus? Carolina Bento explores LDA through a practical real-estate example.
towardsdatascience.com/linea…
9
2,241
With the arrival of Jev, practitioners can now choose between generation-focused models and decision-making-based ones.
@nhu_hoang offers a thoroughly tested comparison of both approaches in her excellent technical deep dive. towardsdatascience.com/jev-v…
1
9
3,020
"I built three versions: retrieval only, a deterministic action planner, and a hybrid system that connects retrieval with actions. Then I ran the same nine tasks through all three."
@emmimalpa presents a hands-on deep dive on the approach she built for bridging the gap between RAG and AI agents. towardsdatascience.com/rag-i…
2
4
14
3,458
"In any real application, you don't just need the next value. You need a forecast, a prediction over an extended horizon."
Waleed Esmail unpacks autoregressive rollout and uncertainty propagation in his recent deep dive. towardsdatascience.com/your-…
2
5
34
6,130
From quantum computing and cloud architecture to digital sovereignty, @rpdesai24 outlines technological topics she believes data scientists should become more fluent in. towardsdatascience.com/10-th…
1
4
13
5,336
Sometimes, the reliability mechanisms we add to LLM pipelines are the ones that make them confidently wrong.
Hubert García Gordon explains why that happens — and how we can prevent this issue from popping up in the first place. towardsdatascience.com/when-…
3
1
12
5,077
Don't miss part 1 of Gal Arav's excellent series on spec-driven test automation: it outlines in great detail what a Verification & Validation standard might look like for agentic systems. towardsdatascience.com/towar…
5
11
5,094
"But if we want to get down to the nitty-gritty and inspect each claim at its most fundamental level, we need a new unit of control: the atomic claim."
Ari Joury explains what it takes to built AI systems that move beyond RAG's inherent limitations. towardsdatascience.com/beyon…
1
4
12
4,670
Is an AI that can prove its own claims within reach? Ari Joury believes we can get there by building a system that goes beyond document-based retrieval. towardsdatascience.com/beyon…
1
2
12
4,252
If it feels like you're going through your coding-agent subscription credits very (too?) quickly, you'll want to explore the recommendations @EivindKjos offers in his helpful guide. towardsdatascience.com/how-t…
10
4,713
Probability can look simple until you ask what the numbers actually mean. Spyros Georgopoulos breaks down Bayesian and frequentist reasoning and connects the concepts to a real marketing mix model.
towardsdatascience.com/why-y…
2
3
22
6,429
“Just use an LLM” is an easy answer, but it may not always be the most efficient one. Himanshu Sharma tests a decades-old baseline to see how far it can go with limited labeled data.
towardsdatascience.com/how-m…
3
4
19
6,626
What changes when an AI agent is designed as a graph instead of a simple loop? Nhu Hoang explores the ideas behind graph engineering and how builders can apply them to real workflows.
towardsdatascience.com/graph…
1
5
23
5,227
Coordinating agent capability development with the application it powers
🖋️ by Andrew Hinton
towardsdatascience.com/where…
2
14
4,474
Giving an LLM permission to act, in 9 steps by Ari Joury, PhD
towardsdatascience.com/how-t…
2
1
8
6,849
Yu Dong replaced a weekly Tableau routine with an agent skill. It inspects the CSV, cross-validates it against the data source, then recommends one chart.
towardsdatascience.com/shipa…
1
1
17
4,818
The model is only part of the equation when working with frontier AI. Eivind Kjosbakken breaks down his early experience with GPT-6 Astra and the techniques he uses to get better results.
towardsdatascience.com/how-t…
2
13
3,438
In a focused new walkthrough, @Himansh70809561 pits autoencoders against PCA to see which performs better in anomaly-detection tasks. towardsdatascience.com/autoe…
3
15
4,448
In a new, hands-on deep dive, Abdullahi Dattijo shows how far you can push a search agent in terms of performance in relation to model calls. towardsdatascience.com/can-a…
1
2
11
4,097