Published inBumble TechBumble Inc open sources Private Detector™ and makes another step towards a safer internet for womenA version of the model powering the feature that detects and blurs lewd images is now available on GithubOct 24, 20222Oct 24, 20222
Published inBumble TechMultilingual message content moderation at scalePart 2: Embeddings space analysis and cross-lingual representationsMay 24, 2022May 24, 2022
Published inBumble TechMultilingual message content moderation at scalePart 1: introduction, model’s design and production infrastructureNov 11, 2021Nov 11, 2021
Published inTDS ArchiveSetting up a CI using dbt Cloud and SnowflakeMerging to master and sleeping at nightOct 7, 2020Oct 7, 2020
Published inTDS ArchiveTesting Airflow jobs on Google Cloud Composer using pytestA reliable CI/CD without reinventing the wheelApr 27, 20201Apr 27, 20201
Something on these crazy timesWorking from home, smart lockdowns and baking powderApr 12, 2020Apr 12, 2020
Published inTDS ArchiveDoing Data in a “normal” companyA personal and honest reviewMar 2, 20201Mar 2, 20201
Published inData Science @ HousingAnywherePredict your rent with Machine LearningBehind RentRadar, the new AI tool by HousingAnywhereNov 28, 2019Nov 28, 2019
Published inTDS ArchiveFrom 0 to Continuously Delivered Machine LearningA step-by-step approach that worked fineNov 4, 2019Nov 4, 2019
Published inTDS ArchiveCoding consciousnessA very long post about the Turing TestSep 20, 2019Sep 20, 2019
Published inTDS ArchiveDetecting scams using AI, for realA year-long journey to develop one of the best fraud prevention systems on the marketJul 22, 2019Jul 22, 2019
If it’s interpretable it’s pretty much useless.Black-box Machine Learning is the only Machine Learning we needApr 30, 201918Apr 30, 201918
Published inTDS ArchiveRandom thoughts on my first ML deployment5 things I didn’t know six months ago and that’s better not to forget in the months to comeMar 3, 20192Mar 3, 20192
Published inData Science @ HousingAnywhereListings similarities using leavesA way to compute similarities among inventory items using pre-trained leaves embeddings.Feb 6, 2019Feb 6, 2019
Published inTDS ArchiveNeural Networks and Philosophy of LanguageWhy Wittgenstein’s theories are the basis of all modern NLPJan 7, 20194Jan 7, 20194
Published inTDS ArchiveXGBoost is not black magicAnd not imputing missing values isn’t always the right choice.Dec 13, 20184Dec 13, 20184
Published inData Science @ HousingAnywhereHow to do Data Science on an evolving WebsiteWhy UX design plays a key role in dataset consistency and how ensemble modeling can help you solve thatNov 30, 20181Nov 30, 20181