Towards Data Science

Advanced Topic Modeling with LLMs | Towards Data Science A deep dive into topic modeling by leveraging representation models and generative AI with BERTopic

Go beyond LDA. Alex Davis shows you how to use representation models + generative AI to unlock deeper insights with BERTopic.

Hands‑On with Agents SDK: Your First API‑Calling Agent | Towards Data Science A practical, beginner‑friendly guide to building an AI weather assistant with Python, OpenAI Agents SDK, API tools, and Streamlit.

Build your own AI assistant! Iqbal Rahmadhan walks you through creating a weather agent using the OpenAI Agents SDK, Streamlit, and Python.

How I Fine-Tuned Granite-Vision 2B to Beat a 90B Model — Insights and Lessons Learned | Towards Data Science A hands-on journey exploring fine-tuning techniques that unlock the power of small vision models.

A hands-on journey exploring fine-tuning techniques that unlock the power of small vision models.

By Julio Sanchez

How To Significantly Enhance LLMs by Leveraging Context Engineering | Towards Data Science The benefits and practical aspects of context engineering for LLMs

Want more out of your LLMs? Eivind Kjosbakken explains how context engineering can dramatically boost performance. From system prompts to strategically feeding the right data.

Torchvista: Building an Interactive Pytorch Visualization Package for Notebooks | Towards Data Science Building a tool to interactively visualize the forward pass of any Pytorch model from within notebooks.

Explore the future of @pytorch.org model visualization with Sachin Hosmani introduces Torchvista, an open-source package that transforms static graphs into interactive visualizations directly in your notebooks.

Decision Trees Natively Handle Categorical Data | Towards Data Science But mean target encoding is their turbocharger

Decision trees can natively handle categorical data, but this ability comes at a computational cost when category counts grow. Read the full article by Vadim Arzamasov now, no paywall.

How to Build an MCQ App | Towards Data Science Key steps to using RAG for generating MCQs from Wikipedia articles based on user-defined context

Reinhard Sellmair explains how to build an app that generates multiple choice questions (MCQs) on any user-defined subject. Learn how the app works, how Wikipedia articles are retrieved, and how these are used to invoke a chat model.

The Journey from Jupyter to Programmer: A Quick-Start Guide | Towards Data Science Explore the real benefits of ditching the notebook

Most Data Scientists start their coding journey using a @jupyter.org Notebook.

Lucy Dickinson shares a clear understanding of how to structure a small data science project as a Python program and appreciate the advantages of transitioning to a programming approach.

MCP Client Development with Streamlit: Build Your AI-Powered Web App | Towards Data Science MCP client development with Streamlit to enhance the tool calling capabilities of remote MCP servers, from setting up your development environment and securing API keys, handling user input,…

Build your own AI-powered web app using MCP and Streamlit. This hands-on guide by Destin Gong walks you through creating an interactive client interface to connect with multiple MCP servers. No heavy frontend required.

New to LLMs? Start Here  | Towards Data Science A guide to Agents, LLMs, RAG, Fine-tuning, LangChain with practical examples to start building

If you are new to LLMs, START HERE 🚩

It can be overwhelming to start studying LLMs with all this content over the internet. But this is the ultimate guide from Alessandra Alpino gives you all the essentials.

Linux Roadmap - roadmap.sh Step by step guide to learn Linux in 2025. We also have resources and short descriptions attached to the roadmap items so you can get everything you want to learn in one place.

Everything you need to go from “what’s a distro?” to Linux pro.

🐧 Linux Roadmap 2025:

– 🖥 Terminal Basics & File System
– 🔧 Permissions & Processes
– 📦 Package Managers & Shell Scripting
– 🔐 Users, Groups, and Security
– 🌐 Networking & Cron Jobs

MCP Client Development with Streamlit: Build Your AI-Powered Web App | Towards Data Science MCP client development with Streamlit to enhance the tool calling capabilities of remote MCP servers, from setting up your development environment and securing API keys, handling user input, connectin...

MCP Client Development with Streamlit: Build Your AI-Powered Web App in @towardsdatascience.com

towardsdatascience.com/mcp-client-d...

Write for Towards Data Science | Towards Data Science Quick Links: Why become a contributor? Earn money with the TDS Author Payment Program Submission Guidelines How to get your article ready for publication! How to submit your work Adding and using…

Want to share your expertise too? Submit your article to be published on TDS.

“I think of analysts as data wizards who help their product teams solve problems” | Towards Data Science Mariya Mansurova explains how hands-on learning, agentic AI, and engineering habits shape her writing and work.

How do you stay ahead in the rapidly evolving world of AI? ✨ Our latest Author Spotlight features Mariya Mansurova, who shares how hands-on learning, agentic AI, and engineering habits shape her impactful work.

Mastering NLP with spaCy – Part 2 | Towards Data Science POS tagging, dependency parser and named entity recognition.

Struggling to make machines truly understand text beyond just words? This new article by Marcello Politi, Part 2 of a series, dives into POS tagging, dependency parsing, and Named Entity Recognition (NER) with spaCy.

How Computers “See” Molecules | Towards Data Science Generative Molecular Design (Part 1): common molecular representations in data science.

How do computers "see" molecules when they only process binary signals? Tianyuan Zheng's debut TDS article explores the fundamental challenge of representing molecules for AI, from simple chemical formulas to complex graph structures.

When Models Stop Listening: How Feature Collapse Quietly Erodes Machine Learning Systems | Towards Data Science Models don't just fail with noise; they fail in silence, by narrowing their attention to the point of fragility.

Your ML model is stable, but silently failing in production? 😫 Mahe Jabeen Abdul's new article exposes feature collapse, where models narrow their attention to a few features, becoming brittle and unreliable.

Not Everything Needs Automation: 5 Practical AI Agents That Deliver Enterprise Value | Towards Data Science What actually works with AI agents inside enterprise organizations?

What actually works with AI agents inside enterprise organizations?

Weiwei Hu shares the top 5 use cases that matter most if you’re looking for clarity on how to start or scale your enterprise AI journey.

Things I Wish I Had Known Before Starting ML | Towards Data Science Part 1: Data, Sales Pitches, Bugs, and Breakthroughs

ML isn't all breakthroughs and papers. It's also sales decks, data messes, and debugging marathons. Pascal Janetzky shares what he wishes he’d known before starting his journey.

MCP Client Development with Streamlit: Build Your AI-Powered Web App | Towards Data Science MCP client development with Streamlit to enhance the tool calling capabilities of remote MCP servers, from setting up your development environment and securing API keys, handling user input,…

Learn how to create a clean and powerful MCP client using Streamlit. Destin Gong shows how to leverage the Model Context Protocol for building interactive AI apps with minimal setup and maximum flexibility.

Developer Roadmaps - roadmap.sh Community driven roadmaps, articles and guides for developers to grow in their career.

How well do you REALLY know that LLM architecture?? 🎓 Test it with a custom quiz from ours sister site @roadmapsh.bsky.social's AI Tutor! Pick the topic, choose the question type, and let AI challenge your understanding.

How To Significantly Enhance LLMs by Leveraging Context Engineering | Towards Data Science The benefits and practical aspects of context engineering for LLMs

Context is everything. Eivind Kjosbakken explores how structured inputs, context-aware prompts, and developer-driven strategies can take LLMs from good to great. Dive into the principles of effective context engineering.

FastSAM  for Image Segmentation Tasks — Explained Simply | Towards Data Science Image segmentation is a popular task in computer vision, with the goal of partitioning an input image into multiple regions, where each region represents a separate object. Several classic approaches…

Your image segmentation model is too slow!! Vyacheslav Efimov's latest article introduces FastSAM, a revolutionary model that's 50x faster than SAM while maintaining high segmentation quality. Learn its architecture and how it achieves this speed.

How to Benchmark LLMs – ARC AGI 3 | Towards Data Science Learn how to LLMs are benchmarked, and try out the newly released ARC AGI 3

Struggling to compare frontier LLMs effectively? Eivind Kjosbakken's latest article dives into the newly released ARC AGI 3 benchmark and explains why even top models score 0%. Get insights into the current limits of AI intelligence.

LLMs and Mental Health | Towards Data Science Are LLMs good or bad for our mental health? It’s more complicated than that.

Are general-purpose LLMs safe for sensitive mental health conversations? @StephanieKirmer.com's new deep dive reveals why non-fine-tuned models can fall short, sometimes with alarming or even harmful consequences.

Stuck on a tricky concept?? 🤔 Our AI Tutor is your new superpower! Just enter any topic you want to master, and get an AI-generated course, customized quizzes, and chat support when you need it.

▶️ roadmap.sh/ai

The ONLY Data Science Roadmap You Need to Get a Job | Towards Data Science Are you looking to become a data scientist and don’t know where to start? In this article, I want to provide you with a straightforward, no-nonsense learning roadmap that you can follow to break into…

Struggling to know where to start in data science? 💡 Egor Howell's latest roadmap is designed to reduce overwhelm and accelerate your job search! Discover the essential skills in statistics, math, programming, and machine learning that truly get you hired.

Getting AI Discovery Right | Towards Data Science A guide to ideating, validating, and prioritizing your AI use cases

How do you build AI solutions that ACTUALLY get adopted? Dr. Janna Lipenkova reveals how to navigate the AI solution space effectively, define strategic opportunities, and foster iterative design for optimal system development.

Eager to amplify your data science expertise? Turn your Machine Learning, AI, or Data Engineering insights into a high-impact article on Towards Data Science and become a recognized voice in the field.

Submit your article today ➡️ bit.ly/TDSContributor

The future of data engagement is here, powered by AI. ✨ Our new eBook unveils how intelligent data tools amplify human insight and democratize building. With support from @plotly.com, download your FREE copy NOW 👉 bit.ly/ai-data-apps...