MissingLink

Deep learning automation for engineers

get it

MissingLink helps data scientists and engineers streamline and automate the entire deep learning cycle. It eliminates grunt work associated with the machine learning process and accelerates the time it takes to train and deliver effective models.

Reviews
  • Tor
    TorCTO, Founder @ Nanit
    Pros: 

    - was instrumental for scaling our AI experiments

    - support makes you feel that you are their only customer

    - constantly adding features

    Cons: 

    - pricing is not ideal, I would have preferred pricing based on number of objects and not size of objects but it’s minor.

    Missing link solutions have helped us build our experiments database from the ground up and saved us weeks to months of work, and let our team focus on building our product.

    Tor has used this product for one year.
  • Pros: 

    If you are training models on different data sets, this solution is the end to mess you get in your data and experiments

    Cons: 

    .,...

    I've seen this guys work on the product for the last year and what it does for their customers. You should definitely try it if you are handling tons of data in many experiments training different models.

    Avi Barliya has never used this product.
Discussion
You need to become a Contributor to join the discussion.
Yosi Taguri
Yosi TaguriMaker@yosi_taguri
Thanks for hunting us, @ryanlawler. Hi PH community, we’re very excited to publicly launch MissingLink.ai, a product we’ve been working on for almost 2 years and which solves a problem we’ve experienced personally as practitioners of deep learning. What’s the problem? Deep learning promises to impact lives in profound ways, but training and running a deep learning model is complex, resource-intensive, and time-consuming. It requires managing multiple components and versions of those components, including the raw data, completed experiments, compute resources, and code. This also makes it extremely difficult to reproduce experiments, which is critical in training effective models. What’s our solution? MissingLink automates the entire deep learning cycle: data, code, experiments, and resources. It automates repetitive time-consuming tasks, shortening model training times. MissingLink provides a versioned petabyte-scale data store, enables organizations to manage hybrid resources (cloud & on-premise) as a single pool with one command, and can be easily integrated with their existing code, data, and infrastructure. Our early customers were able to speed up their experiments by 20x. So if you’re working on a deep learning project, give it a try (there’s a free account). And we’d love to hear your feedback.
Uri Eliabayev
Uri Eliabayev@urieli17 · Foundr, AutoHash
Can't wait to try it. It truly can save me so many work hours!
Eric Nakagawa
Eric Nakagawa@ericnakagawa1
I recently met up with @yosi_taguri to learn about Missing Link. Missing Link helps Data Scientists spend less of their time managing their infrastructure and more of their time experimenting. My current work has me painfully maintaining relatively simple AI infrastructure. This service is a no-brainer for any data science team trying to experiment, iterate, and learn as quickly as possible.
Vadim Solovey
Vadim Solovey@vadimska · CTO
Since I know the team behind MissingLink, the product is going to be a hit!
Niv Calderon
Niv Calderon@niv_calderon · Investor Relations @ Envoy There
@yosi_taguri is one of these rare founders you want to follow wherever he goes and whatever he does, period.