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
Became a MC with TensorFlow
Search
Igor Halfeld
July 19, 2019
Programming
60
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Became a MC with TensorFlow
Training tensorflow with funks to produce funk letters
Igor Halfeld
July 19, 2019
More Decks by Igor Halfeld
See All by Igor Halfeld
Modular monoliths with fastify
igorhalfeld
0
210
How to migrate legacy frontend?
igorhalfeld
1
140
Patterns for perfect components in Vue.js
igorhalfeld
0
140
Clean Architecture no Front-end (update 2021-05-11)
igorhalfeld
0
76
Step by step I followed a career in software
igorhalfeld
0
710
Using AI to create memes
igorhalfeld
0
540
Deninho, the TS children
igorhalfeld
1
230
Micro Front-ends, what they don't tell you
igorhalfeld
0
410
CSS in JavaScripto, the anti-pattern that is good
igorhalfeld
1
110
Other Decks in Programming
See All in Programming
tsc.rip を支える技術 / Kyoto.なんか #8
susisu
0
4.3k
新人はどこまで自力でやり、どこからAIに頼るべきか/エンジニア育成に向き合う_先輩たちの悩みと知見共有会
toppan_digital_dev
1
560
AWS Step Functions 大規模並列の壁を越える / jaws-sonic-2026-niigata-step-functions
kasacchiful
PRO
0
360
iOS開発×AI駆動開発 〜最近使って便利だったスキルの話〜
nogu66
0
140
Claude Codeを組織的に動かして月400PRを実現した話
happy_ryo
0
250
Go 1.27からのGODEBUG / Go 1.27 リリースパーティ #go127party
mazrean
0
310
書籍「プロフェッショナルAI駆動開発」紹介スライド
juntaromatsumoto
0
990
{ Android | Kotlin } Gradle Plugin in 2026
ryunen344
1
280
私のClaude Code活用法 (個人開発編) - PHPerKaigi mini #4(2026/08/24)
panda_program
1
210
信頼性の目標を誰も求めてない
shubox
0
490
Flow は今どうなっているか
mizdra
PRO
0
920
思考垂れ流し開発 ~音声入力 × AIエージェント × 開発ハーネスによる試行錯誤~
npostring
0
810
Featured
See All Featured
Building Better People: How to give real-time feedback that sticks.
wjessup
370
20k
Bioeconomy Workshop: Dr. Julius Ecuru, Opportunities for a Bioeconomy in West Africa
akademiya2063
PRO
1
340
Six Lessons from altMBA
skipperchong
29
4.5k
Groundhog Day: Seeking Process in Gaming for Health
codingconduct
0
330
Creating an realtime collaboration tool: Agile Flush - .NET Oxford
marcduiker
35
2.6k
A designer walks into a library…
pauljervisheath
211
25k
Into the Great Unknown - MozCon
thekraken
41
2.7k
Odyssey Design
rkendrick25
PRO
2
800
Google's AI Overviews - The New Search
badams
0
1.6k
What Being in a Rock Band Can Teach Us About Real World SEO
427marketing
0
1.1k
Documentation Writing (for coders)
carmenintech
77
5.5k
[SF Ruby Conf 2025] Rails X
palkan
2
1.3k
Transcript
Virando um MC com TensorFlow...
Letícia S. Pedroso Analista de Inteligência Artificial - Mutant Instrutora
de Lógica de Programação - Sesc SP @developers-br, @nerdzao @leticiaspedroso - @pedrososleticia Igor L. Halfeld Microsoft MVP Sr. Software Engineer - Estratégia Concursos @vuejs_brasil, @vuejssp @jssaopaulo, @nodebr @igorhalfeld
De onde veio essa ideia? Todo mundo quer ser MC
Vamos primeiros pra alguns conceitos...
Natural Language Processing Processamento de Linguagem Natural
Natural Language Processing é a interseção entre... Ciências da computação
Inteligência artificial Linguística
Corretores...
Tradutores...
Chatbots Assistentes virtuais Conversas...
NER - Named Entity Recognition
Classificadores de conteúdos
Speech Recognition API
None
Funk Generator
Scraping - Puppeteer Treinamento/Geração - TensorFlow Front-end - Vue.js API
- Azure Functions + MongoDb 3 grandes partes pro projeto acontecer...
1000 Letras de música do site letras.com 1 - Scrapping
None
None
None
Pegar as 1000 letras fazer o treinamento e mandar pra
API pra gravar no banco. 2 - Treinamento/Geração
Palavrões que não existem Frequência maior de algumas palavras Mistura
de idiomas palavras em inglês e nome de marcas Alguns padrões encontrados...
Prever o próximo carácter/palavra a partir do estado anterior e/ou
futuro. Ex: calça a bota e bota a calça Tipo de rede escolhida Recurrent Neural Network -RNN
Usar o Vue.js pra construir a interface pegando os funks
da API e usando o speech recognition 3 - Front-end
None
Azure function junto com MongoDB 3 - API
None
None
O funk tem o seu idioma próprio e queremos entende-lo!
Por que usar o tensorflow pra isso?
BORA PRA DEMO!
Obrigado! @igorhalfeld - @pedrososleticia