Greg Reda

@gjreda

machine learning • led data science & • baseball, bikes, and music • Chicagoan

San Francisco, CA
Vrijeme pridruživanja: lipanj 2011.

Tweetovi

Blokirali ste korisnika/cu @gjreda

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @gjreda

  1. Prikvačeni tweet
    8. kol 2019.

    👋 I've been checking Twitter less frequently. My email's on my website if you'd like to reach me.

    Poništi
  2. 13. pro 2019.

    Congratulations to everyone ! What a great accomplishment from an even better group of people! I'm honored to have been part of your journey

    Poništi
  3. 5. lis 2019.

    Welp, first earthquake under my belt now. What a weird feeling.

    Poništi
  4. 2. srp 2019.

    We've reached peak "we're just getting started."

    Poništi
  5. 14. lip 2019.

    As someone using these to prep for interviews, they’re well worth the $12 price

    Poništi
  6. proslijedio/la je Tweet
    8. lip 2019.

    Selecting features using all data before splitting into folds for training/testing is a big source of train-test leakage. To demonstrate, I generated random data and labels, select down to 25 features, and train a model. Much better than random performance due to the leakage.

    Prikaži ovu nit
    Poništi
  7. 1. lip 2019.

    9 days, 10 states, and 2900+ miles later.

    Prikaži ovu nit
    Poništi
  8. 26. svi 2019.

    From the top of Bighorn National Forest, looking west. Think that’s Yellowstone in the distance.

    Prikaži ovu nit
    Poništi
  9. 26. svi 2019.

    The Black Hills are a very beautiful drive and the vast openness of Wyoming is stunning.

    Prikaži ovu nit
    Poništi
  10. 21. svi 2019.

    And if you're in Chicago and looking, look no further than . The leadership is inspiring, the team cares deeply, and everyone truly strives to embody their values () Thank you , , and my old team for a great 3.5 years.

    Prikaži ovu nit
    Poništi
  11. 21. svi 2019.

    Some news: I'm moving to San Francisco and on the job market. A new adventure sounded fun and my wife got a great opportunity. Road tripping over the next 10 days, but if you're doing DS / ML / product stuff in SF, I'd love to buy you coffee once we get there. DMs are open.

    Prikaži ovu nit
    Poništi
  12. 2. svi 2019.

    Really enjoyed this talk on Representation Learning:

    Poništi
  13. 12. tra 2019.

    And to clarify, we're talking your standard chocolate brownie.

    Prikaži ovu nit
    Poništi
  14. 12. tra 2019.

    Settle a team argument: Brownies are ...

    Prikaži ovu nit
    Poništi
  15. 29. lis 2018.

    When screening candidates for technical roles, I'm partial to short take-home challenges over live programming in person or over hangout. Take-home challenges let candidates work how they are most comfortable. Live programming feels pressure-filled and not how work should be.

    Poništi
  16. 9. lis 2018.

    I've been using pandas for over 6 years. My tutorial on it has been viewed over 700,000 times. I still don't know how to use a MultiIndex and would be lost without Google and Stack Overflow. It's ok for you to feel the same.

    Poništi
  17. 5. lis 2018.
    Odgovor korisniku/ci

    This is a good thread for those seeking to enter data science. I agree with very much. Much of a data scientist’s role - the softer skills and tech/quant nuance - come over time, with experience. Staring as an Analyst is a great place to develop those skills.

    Poništi
  18. 6. ruj 2018.

    $1B idea: someone build a search platform that hooks into gmail, slack, dropbox, gdrive, etc. Far too much "where have I seen this before?" and then tons of time finding.

    Poništi
  19. proslijedio/la je Tweet
    3. ruj 2018.
    Odgovor korisniku/ci

    One of the most unintuitive things about ML is that often one good feature, in a (basically) linear model gets you 90+% of the benefits. We spend so much time learning a portfolio of techniques/tricks/tips and heuristics for when to use, then it's mostly about what data you have.

    Poništi
  20. proslijedio/la je Tweet
    9. kol 2018.

    Some advice for people getting into data science/machine learning/AI/etc: metrics and evaluation are the most important thing. Know how to evaluate your model and how to measure and explain its performance over time. Know in which situations it does well or poorly, and why.

    Prikaži ovu nit
    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·