Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Data & Machine Learning at 90 Seconds
Search
Sponsored
·
Ship Features Fearlessly
Turn features on and off without deploys. Used by thousands of Ruby developers.
→
Dat Le
February 26, 2019
Technology
0
100
Data & Machine Learning at 90 Seconds
BigData.SG & Hadoop.SG meetup - Feb 26, 2019
Dat Le
February 26, 2019
Tweet
Share
More Decks by Dat Le
See All by Dat Le
Building your data science capabilities - Nulab Drinking Code 2018-08-31
lenguyenthedat
0
120
Building your data science capabilities - Asia InsurTech 2018-04-23
lenguyenthedat
0
550
Data Science at honestbee - DSSG 2016-10-24
lenguyenthedat
5
2.6k
Data Analytics Infrastructure - DSSG 2015-11-23
lenguyenthedat
2
430
Rakuten - Viki Data Challenge Solution.
lenguyenthedat
0
700
Data Infrastructure with Amazon Web Services
lenguyenthedat
1
200
Other Decks in Technology
See All in Technology
組織のSREを推進するためのPlatform EngineeringとEKS / Platform Engineering and EKS to drive SRE in your organization
chmikata
0
190
Security Diaries of an Open Source IAM
ahus1
0
210
OpenClawで回す組織運営
jacopen
3
640
組織全体で実現する標準監視設計
yuobayashi
1
160
【SLO】"多様な期待値" と向き合ってみた
z63d
2
320
事例に見るスマートファクトリーへの道筋〜工場データをAI Readyにする実践ステップ〜
hamadakoji
0
230
Claude Codeの進化と各機能の活かし方
oikon48
20
9.6k
OCI Security サービス 概要
oracle4engineer
PRO
2
13k
Claude Cowork Plugins を読む - Skills駆動型業務エージェント設計の実像と構造
knishioka
0
300
kintone開発のプラットフォームエンジニアの紹介
cybozuinsideout
PRO
0
840
ナレッジワーク IT情報系キャリア研究セッション資料(情報処理学会 第88回全国大会 )
kworkdev
PRO
0
100
Exadata Database Service on Dedicated Infrastructure(ExaDB-D) UI スクリーン・キャプチャ集
oracle4engineer
PRO
8
7.1k
Featured
See All Featured
[RailsConf 2023] Rails as a piece of cake
palkan
59
6.4k
Creating an realtime collaboration tool: Agile Flush - .NET Oxford
marcduiker
35
2.4k
Being A Developer After 40
akosma
91
590k
Git: the NoSQL Database
bkeepers
PRO
432
66k
Reflections from 52 weeks, 52 projects
jeffersonlam
356
21k
The Myth of the Modular Monolith - Day 2 Keynote - Rails World 2024
eileencodes
26
3.4k
The Art of Programming - Codeland 2020
erikaheidi
57
14k
Bootstrapping a Software Product
garrettdimon
PRO
307
120k
Utilizing Notion as your number one productivity tool
mfonobong
4
250
Technical Leadership for Architectural Decision Making
baasie
3
280
The browser strikes back
jonoalderson
0
770
Reality Check: Gamification 10 Years Later
codingconduct
0
2k
Transcript
DATA AND MACHINE LEARNING AT 90 SECONDS @lenguyenthedat . Director
- Data Science & Engineering at 90 Seconds BigData.SG & Hadoop.SG meetup @ AWS
90 Seconds in 90 seconds 1. INTRODUCTION
Cloud Video Creation Platform Backed by Sequoia since 2016 Trusted
by the world’s biggest brands
90 Seconds is a team of over 180 people from
18 nationalities across 7 global bases, including Singapore, Auckland, Sydney, Tokyo, London, Berlin and San Francisco. Where we are located
2. DATA AT 90 SECONDS
THE DATA TEAM AT 90 SECONDS Engineering Data Marketing Finance
Product Talent Sales Operation Customer Experience
THE DATA TEAM AT 90 SECONDS Data Engineer Data Analyst
Machine Learning Engineer DATA WAREHOUSE DATA PIPELINES INFRASTRUCTURE INTEGRATIONS MACHINE LEARNING, DATA PIPELINES INTEGRATIONS BUSINESS INTELLIGENCE, DATA PIPELINES, VISUALIZATION, INSIGHTS
THE DATA STACK AT 90 SECONDS
None
3. CASE STUDY: VIDEO SEARCH ENGINE
CASE STUDY: VIDEO SEARCH ENGINE Videos in, Data out! We
have produced 30,000 videos till date
CASE STUDY: VIDEO SEARCH ENGINE WHAT IF THERE IS A
SEARCH ENGINE THAT ALLOWS INTERNAL STAFF TO SEARCH FOR VIDEOS BY THEIR ENTITIES?
CASE STUDY: VIDEO SEARCH ENGINE
CASE STUDY: VIDEO SEARCH ENGINE
CASE STUDY: VIDEO SEARCH ENGINE
CASE STUDY: VIDEO SEARCH ENGINE
CASE STUDY: VIDEO SEARCH ENGINE
4. FOOD FOR THOUGHT: HOW LONG DID IT TAKE
FOR YOUR MACHINE LEARNING MODEL TO GO LIVE ON PRODUCTION? one of the most important metric of a successful data team
1. FIND THE RIGHT TALENTS AND TEAM COMPOSITION
2. DATA INFRASTRUCTURE. BUILD IT FIRST!
3. COMMUNICATION & PRIORITIZATION BUY-IN AND UPWARDS MANAGEMENT
4. DO NOT OVERCOOK. GO FOR AN MVP AND ITERATE!
THANK YOU!