mirsci @mirsci

Toronto, Ontario Joined February 2015

Tweets

81
Followers

7
Following

85
Likes

276

KSE @semanticbeeng

a year ago

@Camunda @viadee_it @BenjWolters @viadeeJava x.com/ChristopherA/s… Considering example app: ML workflow where Gordian envelopes are used by a process to orchestrate & monetize #datacontract based access across #dataspaces with sensitive data #bpmn #datamanagement #MLWorkflow (aka "data pipelines") Thoughts? 🤔

Christopher Allen @ChristopherA

2 years ago

1 0 2 143 0

Download Image

1 1 1 72 0

KSE @semanticbeeng

5 years ago

#cloudflow orchestration has nice in-memory semantics which allow sharing a #ApacheSpark session between stages. Not the case for #MlFlow, for example (must recreate session)

Akka @akka_io_

5 years ago

#cloudflow orchestration has nice in-memory semantics which allow sharing a #ApacheSpark session between stages. Not the case for #MlFlow, for example (must recreate session) https://t.co/cVEXAXQ3aU

0 1 2 0 0

Download Image

0 2 3 0 0

Download Image

KSE @semanticbeeng

6 years ago

"Why not Airflow?" "#Airflow popularized many of the workflow semantics #dataengineer -s take for granted- fails to meet more dynamic needs as #dataengineering discipline matures" "does not support #dataflow in a first class way" #MachineLearningEngineering #WorkflowManagement

1 2 4 0 0

Download Image

KSE @semanticbeeng

5 years ago

#Dagster ⇨ "data dependencies, not with pure execution dependencies" ⇨ "structured #ProgrammingModel w/ richer, semantically aware graph" #ProgrammingModel abstractions over #DataOps & #MLOps #Dataflow + #ControlFlow #MLWorkflow #StatefulApps #MachineLearningEngineering

1 4 4 0 1

Download Image

mirsci @mirsci

6 years ago

#DataPrivacy analysis in the context of #GDPR compliance implemented by @XAIN_AG, as outlined in xain.io/assets/XAIN-Wh… using #FedML #MachineLearning, and having as main goals:

0 1 1 0 0

Download Image

mirsci @mirsci

7 years ago

Design for securing an outsourced and #distributed #MachineLearning (as a service) system based on #SGX enclaves, #Ryoan sandboxes, #Theano, from @rzshokri , Congzheng Song, Tyler Hunt, Vitaly Shmatikov, Emmett Witchel

Software Engineering @ComputerPapers

8 years ago

0 1 1 0 0

0 0 1 0 0

mirsci @mirsci

7 years ago

#Google released Feast, its new open source feature store to manage storage and access of #MachineLearning features github.com/gojek/feast/

Jesus Rodriguez @jrdothoughts

7 years ago

#Google released Feast, its new open source feature store to manage storage and access of #MachineLearning features github.com/gojek/feast/

0 1 2 0 0

0 0 1 0 0

statisticallyfit @anamariavintila

7 years ago

Auto differentiation in Scala code from WinVector: github.com/WinVector/Auto…

0 1 1 0 0

mirsci @mirsci

7 years ago

Focus on building first the foundation for data ingestion, #dataflow, transformation and only then on overlaying complex AI models is critical in building sustainable AI products - “The AI Hierarchy of Needs” by @mrogati hackernoon.com/the-ai-hierarc…

0 2 5 0 1

Download Image

mirsci @mirsci

7 years ago

Just completed a small POC with Karma #karmausc, led by Pedro Szekely (@szeke) for #semantic data mappings. Does anyone know if Karma platform has a support model for production implementations? Thanks.

0 0 1 0 0

Paweł Szulc @EncodePanda

7 years ago

Here's a pattern: yet another company I work for is 100% remote and guess what: they have zero issues finding talented people willing to work for them. Send this tweet to the HR department in your company.

5 50 130 0 1

statisticallyfit @anamariavintila

7 years ago

Just completed a nice function to return symbolic natural cubic spline polynomials in a vector in Matlab.

0 1 1 0 0

Jonas Bonér @jboner

7 years ago

How to kill your Ego. ryanholiday.net/25-ways-to-kil…

0 8 18 0 1

Adrian Colyer @adriancolyer

7 years ago

Serialization and deserialization is a major performance killer in frameworks like Spark and Flink. Skyway bypasses all that and does direct streaming copies from heap to heap across nodes... blog.acolyer.org/2018/04/26/sky…

2 37 106 0 5

mirsci @mirsci

8 years ago

One more reason to embed #functionalprogramming in the #blockchain designs. Excellent post.

0 0 6 0 0

mirsci @mirsci

8 years ago

Scalable #Blockchain platform for enterprise adoption using sharding

Adrian Colyer @adriancolyer

8 years ago

Scalable #Blockchain platform for enterprise adoption using sharding

0 20 57 0 0

Download Image

0 0 0 0 0

Axionable AI @AxionableAI

8 years ago

5 Heroic Tools for Natural Language Processing #NLP 1- CoreNLP from #Stanford 2- NLTK library for #Python 3- TextBlob as an interface 4- Gensim for similarity analysis 5- SpaCy to boast performance bit.ly/2EplsIo #DataScience #chatbot #TextMining

3 43 93 0 2

Download Image

Martin Krasser @mrt1nz

8 years ago

I wrote a brief summary of courses and books that helped me getting started with (and further into) #MachineLearning #DeepLearning and #BayesianStatistics krasserm.github.io/2018/01/03/mac…

0 18 56 0 0

RChain Cooperative @rchain_coop

8 years ago

"Greg has spent many years developing a particularly powerful variation of a mobile process calculus called ρ-calculus (#Rho-calculus), which closes the loop on several issues that are outstanding in a more traditional framework, π-calculus." bit.ly/RChainSeriesIn… @coinfund_io

0 5 10 0 0

Rich Rogers @RichRogers_

8 years ago

It can take a year of programming on a project before you understand what the best design approach should've been. - @martinfowler martinfowler.com/bliki/Technica…