HNHacker News
TopNewBestAskShowJobs

vtuulos

780 karma · joined February 28, 2011

Helping companies build real-world ML/AI systems.

CEO/co-founder/customer support representative at https://outerbounds.com

I started https://metaflow.org at Netflix where I led ML infra for 3+ years. I wrote a book about my learnings:

https://www.manning.com/books/effective-data-science-infrastructure

Earlier projects:

http://traildb.io

http://tuulos.github.io/sf-python-meetup-sep-2013/

http://discoproject.org

submissionscomments
vtuulos··on Luigi vs. Airflow vs. Pinball
We have been using Luigi in production for a year now at AdRoll, to manage a graph of tens of data processing tasks. We have been really happy with it.

You can read more about our setup in these two blog posts:

http://tech.adroll.com/blog/data/2015/09/22/data-pipelines-d...

http://tech.adroll.com/blog/data/2015/10/15/luigi.html

vtuulos··on Ask HN: Who is hiring? (February 2016)
AdRoll - San Francisco

If you like languages such as Python, JS, C, D, Lua, Erlang, AWS, and/or petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

To give an idea how we work, recently we launched a new product on top of a novel data processing pipeline on AWS using Docker, read more about it here:

http://tech.adroll.com/blog/data/2015/09/22/data-pipelines-d...

and more about our approach to data science, see here

http://tech.adroll.com/blog/data-science/2015/08/25/factoriz...

http://tuulos.github.io/pydata-2014/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF.

vtuulos··on Feeding data to 1000 CPUs – comparison of S3, Google, Azure storage
We are routinely pulling a dataset of hundreds of GBs to 100+ instances (1600+ cores) in parallel. We have never noticed throughput going down with the number of nodes. S3 delivers the maximum throughput of 2-4Gbps / instance very consistently.
vtuulos··on Ask HN: Who is hiring? (January 2016)
AdRoll - San Francisco

If you like languages such as Python, JS, C, D, Lua, Erlang, AWS, and/or petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

To give an idea how we work, recently we launched a new product on top of a novel data processing pipeline on AWS using Docker, read more about it here:

http://tech.adroll.com/blog/data/2015/09/22/data-pipelines-d...

and more about our approach to data science, see here

http://tech.adroll.com/blog/data-science/2015/08/25/factoriz...

http://tuulos.github.io/pydata-2014/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF.

vtuulos··on Ask HN: Who is hiring? (November 2015)
AdRoll - San Francisco

If you like languages such as Python, JS, C, D, Lua, Erlang, AWS, and/or petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

To give an idea how we work, recently we launched a new product on top of a novel data processing pipeline on AWS using Docker, read more about it here:

http://tech.adroll.com/blog/data/2015/09/22/data-pipelines-d...

and more about our approach to data science, see here

http://tech.adroll.com/blog/data-science/2015/08/25/factoriz...

http://tuulos.github.io/pydata-2014/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF.

vtuulos··on Petabyte-Scale Data Pipelines with Docker, Luigi and Elastic Spot Instances
We have a few cases of containers pulling (meta)data from databases. We use Postgres on RDS, so setting up a read-replica is very easy. Containers can safely access these replicas.

Raw data is collected by our bidders and adservers that push data to S3 and Kinesis for real time consumers (http://tech.adroll.com/blog/data/2015/06/26/kinesis.html)

vtuulos··on Ask HN: Who is hiring? (September 2015)
AdRoll - San Francisco

If you like Python, JS, C, D, Lua or Erlang, and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

This time we are looking for Pythonistas in particular! We have very interesting challenges for Python experts, not all of which are about web apps.

To get an idea of things we do, see

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF.

PS. we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

vtuulos··on Ask HN: Who is hiring? (August 2015)
AdRoll - San Francisco

If you like Python, JS, C, D, Lua or Erlang, and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

To get an idea of things we do, see

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF.

PS. we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

vtuulos··on Continuum Analytics Raises $24M Series A for Anaconda Python and PyData
Congrats Continuum!
vtuulos··on Ask HN: Who is hiring? (July 2015)
AdRoll - San Francisco

If you like Python, JS, C, D, Lua or Erlang, and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

To get an idea of things we do, see

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF.

PS. we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

vtuulos··on Comparison of Erlang Runtime System and Java Virtual Machine [pdf]
DiscoDB does exactly what you described: It memory maps a large read-only data structure in RAM. It has an Erlang binding based on NIFs.

http://discodb.readthedocs.org/en/latest/

https://github.com/discoproject/discodb

vtuulos··on Programming in Standard ML (2011) [pdf]
If you are interested in Standard ML, take a look at http://mlton.org/

Mlton is one of the few whole-program optimizing compilers. One time I wrote a DSL in Standard ML to analyze 100GB+ of memory-mapped data.

Thanks to Mlton, the result was very performant. Thanks to SML, I felt very productive and the result was pretty much bug-free. Highly recommended.

vtuulos··on Ask HN: Who is hiring? (June 2015)
AdRoll - San Francisco

If you like Python, JS, C, D, Lua or Erlang, and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

Analytics is so critical to us that a large part of the stack, including our high-performance, in-memory data backend, is developed in-house in Python (Numba), Erlang, and C. Read more here

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF. PS. Btw, we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

vtuulos··on Call me maybe: Aerospike
We (Adroll) have been very happy with DynamoDB for use cases like this. Works fine with ~500B keys, while maintaining very low, and consistent, latencies.
vtuulos··on Ask HN: Who is hiring? (May 2015)
AdRoll - San Francisco

If you like Python, JS, C, D, Lua or Erlang, and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

Analytics is so critical to us that a large part of the stack, including our high-performance, in-memory data backend, is developed in-house in Python (Numba), Erlang, and C. Read more here

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF. PS. Btw, we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

vtuulos··on How does D improve on C++17?
We (AdRoll) use it for performance-sensitive machine learning:

http://tech.adroll.com/blog/data/2014/11/17/d-is-for-data-sc...

vtuulos··on Ask HN: Who is hiring? (April 2015)
AdRoll - San Francisco

If you like Python, JS, C, D, Lua or Erlang, and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

Analytics is so critical to us that a large part of the stack, including our high-performance, in-memory data backend, is developed in-house in Python (Numba), Erlang, and C. Read more here

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF.

PS. Btw, we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

vtuulos··on Ask HN: Who is hiring? (March 2015)
AdRoll - San Francisco

If you like Python, JS, C, D, Lua or Erlang, and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

Analytics is so critical to us that a large part of the stack, including our high-performance, in-memory data backend, is developed in-house in Python (Numba), Erlang, and C. Read more here

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF.

PS. Btw, we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

vtuulos··on Ask HN: Who is hiring? (February 2015)
AdRoll - San Francisco

If you like Python, JS, C, D, Lua or Erlang, and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

Analytics is so critical to us that a large part of the stack, including our high-performance, in-memory data backend, is developed in-house in Python (Numba), Erlang, and C. Read more here

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF.

PS. Btw, we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

vtuulos··on Ask HN: Who is hiring? (January 2015)
AdRoll - San Francisco

If you like Python (or C or D or Erlang) and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

Analytics is so critical to us that a large part of the stack, including our high-performance, in-memory data backend, is developed in-house in Python (Numba), Erlang, and C. Read more here

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville@adroll.com) or over coffee in SF.

PS. We are also hiring Javascript engineers! Btw, we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

vtuulos··on Distributed Python for the Erlang Ecosystem
You can use Disco to execute arbitrary Python tasks or workflows - it is not limited to MapReduce.

Disco was designed for data-intensive workloads. It wouldn't make much sense to use it to supervise Python servers/services, for instance.

vtuulos··on HyperLogLog and MinHash
We recently open-sourced the algorithm described in the blog post here:

https://github.com/AdRoll/cantor

Btw, we are hiring! Feel free to ping me at ville@adroll.com if you find topics like this relevant to your interests.

vtuulos··on Ask HN: Who is hiring? (November 2014)
AdRoll - San Francisco

If you like Python (or C or D or Erlang) and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

Analytics is so critical to us that a large part of the stack, including our high-performance, in-memory data backend, is developed in-house in Python (Numba), Erlang, and C. Read more here

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville.tuulos@adroll.com) or over coffee in SF.

PS. We are also hiring Javascript engineers! And, btw, we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

vtuulos··on Binary Search Eliminates Branch Mispredictions (2012)
This is very cool. Thanks for sharing!
vtuulos··on Vitesse Data: Postgres + LLVM
Our DeliRoll is used for analytics. It doesn't support random updates, it is append-only.

We use bitmap indexes in a number of places internally.

vtuulos··on Vitesse Data: Postgres + LLVM
Thanks! Encoding takes care of regularities - it is all transparent to the user and DBA.
vtuulos··on Vitesse Data: Postgres + LLVM
Yeah, lack of aggregation function and join pushdown is annoying.

Our data model makes sure that we don't have to do huge joins on the fly, which would be a bad idea anyways. We have a workaround to distribute medium-scale joins that occur frequently. Small joins are handled fine by Postgres as usual.

I would be curious to learn more about Vitesse. I couldn't find many technical details besides the mailing list post.

vtuulos··on Vitesse Data: Postgres + LLVM
We have been running a similar setup (Postgres -> Foreign Data Wrappers -> LLVM) at AdRoll for over a year. We keep 100TBs+ of raw data in memory, compressed.

We managed to build our solution mostly in Python(!) using Numba for JIT and a number of compression tricks. More about it here:

http://tuulos.github.io/pydata-2014/#/

https://www.youtube.com/watch?v=xnfnv6WT1Ng

vtuulos··on Ask HN: Who is hiring? (October 2014)
AdRoll - San Francisco

If you like Python (or C or D or Erlang) and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

Analytics is so critical to us that a large part of the stack, including our high-performance, in-memory data backend, is developed in-house in Python (Numba), Erlang, and C. Read more here

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville.tuulos@adroll.com) or over coffee in SF.

PS. We are also hiring Javascript engineers! And, btw, we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

vtuulos··on Ask HN: Who is hiring? (September 2014)
AdRoll - San Francisco

If you like Python (or C or D or Erlang) and petabytes of data, this is your dream job. AdRoll is one of the fastest growing adtech company, already producing $100M+ in revenue with a small engineering team.

Analytics is so critical to us that a large part of the stack, including our high-performance, in-memory data backend, is developed in-house in Python (Numba), Erlang, and C. Read more here

http://tuulos.github.io/pydata-2014/

http://tuulos.github.io/sf-python-meetup-sep-2013/

I'm happy to tell you more by email (ville.tuulos@adroll.com) or over coffee in SF.

PS. We are also hiring Javascript engineers! And, btw, we write our own convex optimizers, if machine learning is close to your heart - ping me for more info.

← PreviousPage 4 of 6Next →