Tanya Joshi’s Post

Choosing the best machine learning algorithm isn’t just about accuracy—efficiency matters! 🚀 Check out this quick-reference chart summarizing the time complexity of 10 essential ML algorithms, for both training and inference. From Linear Regression and SVMs to Random Forests and k-Nearest Neighbors, this visual guide shows how each model scales as your data grows—helping you make smarter choices for real-world deployments. Whether optimizing for big data or tight production timelines, understanding these computational costs is key for every ML engineer and data scientist. Which algorithm surprised you most? Let’s discuss! 👇 #MachineLearning #DataScience #MLAlgorithms #BigData #AIEfficiency

For anyone looking to upskill in the AI/ML space, consistent, hands-on practice is essential for mastering the fundamentals. I discovered a fantastic website for this on X a few days ago called Deep-ML.com. After using it for a couple of days, I've found it very much valuable. Think of it as a LeetCode specifically for Machine Learning and Data Science. It offers a curated collection of problems perfect for building and testing your skills across a wide range of categories. As a free tool for interview prep or simply honing your craft, it's excellent. I'm already finding it helpful for strengthening my own basics and highly recommend checking it out and hope it adds value to you

🎉 Day 67 of My Data Science Journey! 🎉 Today was an incredible deep dive into the very logic of uncertainty: probability! 🧠 I went beyond simple chance and learned the foundational principles that allow us to count, reason, and make predictions under uncertainty, which is a core skill for any data scientist. 📈 Key Takeaways & Skills : 1} The Foundations of Counting: I started with set theory and mastered permutations and combinations, which are the essential tools for counting the number of possible arrangements and selections of data. 2} Conditional Probability: I tackled conditional probability, understanding the likelihood of an event occurring given that another event has already happened. This is a game-changer for building predictive models. 3} The Power of Bayes's Theorem: This was the biggest breakthrough! I learned how Bayes's Theorem gives us a formal, mathematical way to update our beliefs and predictions based on new evidence. It's the "why" behind so many machine learning algorithms. What I Found Interesting : It's mind-blowing how a single equation can allow you to reason with uncertainty so effectively. The idea that you can start with a prior belief and refine it with new data using Bayes's Theorem is the very essence of how many AI systems learn and adapt in the real world. Upcoming Learning : With these foundational concepts in probability now in my toolkit, my next step is to apply them to different probability distributions and see how they are used to solve real-world problems. Let's Connect: What's your favorite real-world application of conditional probability or Bayes's Theorem? I'd love to hear some examples! #DataScience #Statistics #Probability #BayesTheorem #LearningJourney #Day67 #DataAnalytics

🚨 Scheduling your Jupyter notebook with cron is a MESS And one better way to do it instead, but let's talk about the mess. - First you might not know how to run cron. - And you need to transform your notebook into something cron can actually run. - You need to hardcode the whole system path context. - Cron might not play nicely with virtual environment or conda, you might need to wrap everything in a bash script. - Maybe your script is not going to work after 5 hours of building it? 😱 ✅ But with Livedocs, you can easily schedule your notebook in 1 click. (example below). No code, no bash script, no virtual environment needed. Read more here 👇 https://lnkd.in/gpr3qH5c Create your notebook for FREE 👇 https://livedocs.com #data #datascience #dataanalytics #ai #notebooks #Jupyter

LLMs are based on Transformers, Transformers on self-attention, and self-attention on the math of separation of concerns. But the job description? Oh, that’s built on chaos theory, merging data science, data engineering, GenAI, ML Engineering, and MLOps into one mythical “unicorn who does everything.” Beautiful how theory and reality never meet. #ArtificialIntelligence #MachineLearning #DataScience #GenAI #LLM #MLOps #AICommunity #TechLeadership #DigitalTransformation #Innovation #FutureOfWork #AITrends #DeepLearning #AIEthics #AITalent

Day 1: Feature Engineering Journey 🚀 Today, I’ve started learning Feature Engineering — an essential step in the Data Science and Machine Learning process. 💡 What I explored today: What exactly is Feature Engineering? Different types of Feature Engineering Simple definitions and practical examples Feature Engineering feels like turning raw data into gold — and I’m just getting started! ✨ 👉 If you’re a Data Scientist or ML enthusiast, what’s one tip you’d give a beginner about Feature Engineering? Drop your thoughts or favorite learning resources in the comments — I’d love to hear from you! 🙌 I’ve also created a note of today’s learning to document my progress. #FeatureEngineering #DataScience #MachineLearning #LearningJourney #AI

Integrate new tabular models into your existing workflows with ease. Dr. Carmen Adriana Martínez Barbosa highlights models like TabPFN and TabDPT that feature a familiar Scikit-learn API for faster adoption.

Lately I’ve been working on unsupervised learning projects, and one lesson keeps standing out: 👉 Feature engineering can make or break your results. Before you even think about clustering or dimensionality reduction, the way you prepare your data completely changes what the algorithm “sees.” A few steps I now treat as non-negotiable: 🔹 Exploring a correlation matrix to understand how features interact — not to remove anything blindly, but to get a feel for the data’s structure. 🔹 Applying normalization when features have very different scales (like income vs age), so each variable contributes fairly to distance calculations. 🔹 Using standardization when methods such as PCA or k-means assume data centered around zero with unit variance. Unsupervised models don’t have labels to guide them — their understanding of the data depends entirely on how we represent it. The cleaner and more balanced the features, the more meaningful the patterns we uncover. #MachineLearning #UnsupervisedLearning #FeatureEngineering #DataScience

📚 One surprise from my Data Science journey so far… I expected the hardest part to be learning algorithms like logistic regression or neural nets. But the real challenge has been: - Turning messy data into something usable - Spotting bias or leakage before it ruins a model - Deciding when “good enough” is actually good enough For example, in one project, my model hit the target metric, but a quick sanity check showed the predictions weren’t realistic. That lesson stuck: accuracy isn’t everything — context matters. 👉 Curious — what’s been the most unexpected challenge in your learning or career journey? #DataDrivenWisdom #BusinessTechMastery #DataScience #LearningJourney #CareerGrowth

A tricky geometry puzzle shows just how much leading LLMs have progressed in little over a year. What took me nearly 2 hours of back and forth using GPT-4o last year, Sonnet 4.5 solves in less than 10 seconds today. I detail everything in my latest article on the Towards Data Science platform. It's completely free to read. Check it out using the link below and see if you can solve the puzzle before looking at the answer! https://lnkd.in/eKgbTrzb