Stay Ahead in the Data Science Game with The Variable Newsletter
Never miss a new edition of The Variable, our weekly newsletter packed with a top-notch selection of editors’ picks, deep dives, community news, and much more. As we wrap up January 2026, it’s currently too early to pinpoint major shifts in industry trends. However, one thing is crystal clear: our readers are eager to stay updated on cutting-edge tools and emerging themes in data science.
- Insightful Contributions from TDS Authors
- The Great Data Closure: Are Databricks and Snowflake at Their Limits?
- LLMs and the Infinite Context Challenge
- Optimizing Claude Code Usage
- Spotlight on January Highlights
- Beyond Prompting: The Power of Context Engineering by Mariya Mansurova
- Cutting LLM Memory by 84% by Ryan Pégoud
- Why Human-Centered Data Analytics is Crucial by Rashi Desai
- Retrieval for Time-Series by Sara Nobrega
- The Supply Chain: A Playground for Data Scientists by Samir Saci
- Federated Learning Basics by Parul Pandey
- Authors in the Spotlight
Insightful Contributions from TDS Authors
At TDS, our contributors have kicked off the year on a strong note, offering timely insights into various topics within the data science landscape. This week, we’re showcasing our most-read and -shared articles from January, including explorations into large language model (LLM) contexts, Claude Code, and the evolving landscape of major data platforms.
The Great Data Closure: Are Databricks and Snowflake at Their Limits?
Hugo Lu’s thought-provoking article, “How big can a data company really grow?” dives into the business models of massive platforms like Databricks and Snowflake. He meticulously unpacks the multiple factors affecting their growth and shares some bold predictions for the upcoming year. This piece raises essential questions about the sustainability of current data practices and is a must-read for anyone vested in data-driven enterprises.
LLMs and the Infinite Context Challenge
In a fascinating exploration, Moulik Gupta addresses whether it really is possible to accomplish more with less in his piece, “How LLMs Handle Infinite Context with Finite Memory.” He introduces the concept of Infini-attention, making complex ideas accessible. For anyone interested in enhancing their understanding of LLMs, this article offers great insights.
Optimizing Claude Code Usage
Eivind Kjosbakken provides a handy guide to maximizing the effectiveness of Claude Code. His article outlines essential optimization techniques for users of this popular agentic-coding tool, making it a vital read for developers looking to leverage Claude Code to its fullest potential.
More Read
Spotlight on January Highlights
In addition to the key articles mentioned above, here are some more highlights from January that cover a range of pertinent topics in the data science realm:
Beyond Prompting: The Power of Context Engineering by Mariya Mansurova
Mariya offers a compelling overview of how to utilize ACE for creating self-improving LLM workflows and structured playbooks. This piece emphasizes the importance of context in optimizing machine learning models.
Cutting LLM Memory by 84% by Ryan Pégoud
In this insightful article, Ryan discusses strategies for resolving out-of-memory (OOM) issues encountered with LLM layers. He explores the implementation of custom Triton kernels, providing valuable solutions for data scientists struggling with memory constraints.
Why Human-Centered Data Analytics is Crucial by Rashi Desai
Rashi makes a strong case for prioritizing human-centered data analytics, illustrating how thoughtful design can deepen our understanding of metrics and enhance the impact of data-driven decisions.
Retrieval for Time-Series by Sara Nobrega
Sara introduces the concept of retrieval in time-series forecasting, detailing how looking back can improve predictions. This foundational knowledge is essential for any data scientist looking to refine their forecasting techniques.
The Supply Chain: A Playground for Data Scientists by Samir Saci
Drawing on a decade of experience, Samir highlights why the supply chain is an ideal domain for data scientists in 2026. He discusses how this field offers a unique opportunity for data professionals to see the value of their skills firsthand.
Federated Learning Basics by Parul Pandey
Parul demystifies the foundations of federated learning in her article, paving the way for those looking to explore this innovative approach to training models where the data resides.
Authors in the Spotlight
We also encourage you to explore our recent author Q&A that showcases the work of some of our newest contributors:
-
Diana Schneider examined evaluation methods for multi-step LLM-generated content, offering new perspectives on customer journeys.
-
Kaixuan Chen and Bo Ma shared their insights on building a neural machine translation system for Dongxiang, a low-resource language, illustrating the importance of inclusivity in AI advancements.
- Pushpak Bhoge provided a thorough performance benchmark of Meta’s SAM 3 compared to specialized models, contributing to a deeper understanding of model capabilities.
If you’re looking to make a statement in the data science community, now is the perfect time to submit your writing for the TDS Author Payment Program.
Subscribe to Our Newsletter
To keep your knowledge updated and enhance your understanding of the rapidly evolving data landscape, don’t forget to subscribe to The Variable. This newsletter is your go-to resource for curated articles, meaningful insights, and opportunities in the world of data science.
Inspired by: Source
