user avatar
Radek Osmulski
@radekosmulski
LLMs and retrieval by day and other genres of AI when I get the chance 🧪 Senior AI Eng @NVIDIAAI 🏫 @fastdotai trained DL Eng 📝 learn-dl.com
see my projects ➝
radekosmulski.com
Joined April 2014
614
Following
29.7K
Followers
Posts
  • Pinned
    user avatar
    Radek Osmulski
    @radekosmulski
    May 31, 2025
    Nearly all the books I read feel life-changing, but this one hit different. You know those concepts that have been almost making sense for years? Several of them just clicked for me while reading this. Going to share excerpts + my thoughts over the next few days & list them 👇
    10K
  • user avatar
    Radek Osmulski
    @radekosmulski
    Sep 18, 2022
    Mathematics for Machine Learning -- a 47-page introduction from UC Berkeley 🚀 • Linear Algebra • Calculus and Optimization • Probability A 100% free resource! Source: gwthomas.github.io/docs/math4ml.p…
  • user avatar
    Radek Osmulski
    @radekosmulski
    Apr 26, 2022
    Today is my first day at @NVIDIAAI! 🥳 -From learning to code at 29 -through learning ML @fastdotai -winning a @kaggle competition -jobs at 🔥 startups -moving continents thx to AI -to joining the illustrious Merlin team ❤️ I am beyond grateful 🙏 Will make this one count!
  • user avatar
    Radek Osmulski
    @radekosmulski
    Dec 26, 2017
    Favorite recent jupyter notebook discovery - the %debug magic: 1. Get an exception. 2. Insert a new cell, type %debug and run it. An interactive debugger will open bringing you to where the exception occurred and allowing you to look around!
  • user avatar
    Radek Osmulski
    @radekosmulski
    Jan 17, 2023
    I just became a @kaggle Grandmaster! 🥳
    128K
  • user avatar
    Radek Osmulski
    @radekosmulski
    Mar 18, 2022
    Ok, this is not something I have expected 🤯 itertuples can be 50 times faster than iterrows!
  • user avatar
    Radek Osmulski
    @radekosmulski
    Sep 28, 2022
    To understand the foundations of NLP (pre-Transformers), where would you go? This 48-page paper is the answer 🤩 ✅ concise and clear explanations ✅ sklearn, spacy, and keras code snippets ✅ all the fundamentals of NLP in a single place papers.ssrn.com/sol3/papers.cf…
  • user avatar
    Radek Osmulski
    @radekosmulski
    Nov 17, 2022
    Merlin Dataloader is 119x faster than my own PyTorch Dataset + Dataloader combo! This is revolutionary for tabular data 🥳 Let's take a closer look at what is going on.
  • user avatar
    Radek Osmulski
    @radekosmulski
    Mar 2, 2019
    This is probably the best intro to probability (along with the associated lectures youtube.com/playlist?list=…), now available online for free ❤
    user avatar
    Joe Blitzstein
    @stat110
    Mar 1, 2019
    New edition of my probability book with Jessica Hwang is out! Read it FREE online at probabilitybook.net
    youtube.com
    Statistics 110: Probability
    Statistics 110 (Probability) has been taught at Harvard University by Joe Blitzstein (Professor of the Practice in Statistics, Harvard University) each year ...
  • user avatar
    Radek Osmulski
    @radekosmulski
    Apr 11, 2023
    How does LangChain actually work? We see the wonderful things it can do, but what does it send to the model? What does the model send back? How does it all work? I decided to investigate 🕵️‍♂️ Here is how LangChain allows LLMs to perform Google searches:
    174K
  • user avatar
    Radek Osmulski
    @radekosmulski
    Feb 24, 2022
    Curious about recommender models? Interested in endowing models from other domains with some of their superpowers? Please join me on a whirlwind tour of 6 recsys architectures! >> a thread 🧵 <<
  • user avatar
    Radek Osmulski
    @radekosmulski
    Sep 28, 2022
    Linear Algebra -- an introductory course to Mathematics at the heart of Machine Learning ✅ 37 bite-sized videos (< 10 minutes) ✅ stellar visualizations ✅ expertly delivered by lecturers from Imperial College London Link: youtube.com/playlist?list=…
  • user avatar
    Radek Osmulski
    @radekosmulski
    Nov 9, 2022
    3 ways to speed up your Python/pandas code by up to 10x that I learned from a recent @kaggle notebook:
  • user avatar
    Radek Osmulski
    @radekosmulski
    Aug 9, 2022
    How to speed up your tabular data processing by 1053x A tutorial on how to vectorize a complex operation in pandas/cudf using a boolean mask Bonus at the end: how to seamlessly run on the GPU with arbitrarily large data 1/19