🎙 talk.js - June 2018
ongmin opened this issue · 9 comments
📝 Details
- Event: talk.js
- Date: TBC (Probably 20th June)
- Hosts: @ongmin
- Venue: Singapore Power, #8-00 Keppel Tower 2, 240 Tanjong Pagar Road, Singapore 088540
📅 Agenda
| Time | Agenda |
|---|---|
| 7.00pm | Doors open - mingle with guests and speakers over 🍕 |
| 7.30pm | Start of event - opening segment by host |
| 7.40pm | Start of scheduled talks |
| 8.40pm | Open announcements - open segment for anyone to make announcements |
| 9.00pm | End of event |
💬 Talks
- TensorFlow.js: Machine learning in the browser by Alvin Chan
- Frontend Dependency Tracing for HTTP/2 Server Push by Sophie Yang
- Automating processes using TagUI tool by Ken Soh
- Building fast and semantic input masks in VueJS by Divya Sasidharan
🌱 Propose a talk this month
Want to speak at this month's talk? Leave a comment below with the following details:
Click for details
Title of your talk
Insert catchy title here
Talk description
Tell us what you'd like to talk about
Presenter(s)
Include GitHub handles where possible. Add your Twitter handle if you wish
Presenter's bio
A short introduction of yourself
Duration requested
10min, 20min, 30min
Deck URL (if any)
You can add this part in later
📢 Open announcements
Have something you'd like to tell the audience? We have an open announcement segment for you to announce job opportunities, promote a meet up, upcoming conferences, etc.
📝 Organizers' checklist
Click for details
- Check available dates on We Build SG
- Confirm date and venue availability
- Create event on SingaporeJS Meetup Page
- Inform @engineersftw for video recordings
- Announce event on SingaporeJS Gitter and other social media platforms
- Prepare deck for event
* checked boxes indicate confirmed/done
Hi there, comments please 😃
Title of your talk
TensorFlow.js: Machine learning in the browser
Talk description
TensorFlow, Google’s machine learning framework has finally arrived in Javascript!
Now you can create, train and run machine models in the browser. This means that data do not need to be transferred to a backend for prediction, and machine learning logic can be combined directly with your frontend magic in Javascript!
This talk will begin with a brief intro about machine learning for some context. Next, we will jump into code samples on how to create, train and run a machine learning model in the browser.
Presenter
Alvin Chan
A short introduction of yourself
Full-stack developer, turned Data Scientist in cybersecurity startup. Starting PhD in Artificial Intelligence (NTU) this August.
Duration
20-30 min (depends on available time)
Title of your talk
Automating processes using TagUI tool
Talk description
TagUI is an automation tool I developed while on a one-year sabbatical in Eastern Europe, after resigning from DBS Technology & Operations (test automation of investment and trading apps).
The tool basically translates human-language-like syntax into JavaScript code that drives UI automation. Thus users can also write JavaScript directly in TagUI scripts. Use cases can be data extraction, process automation, test automation.
As the underlying framework is CasperJS, an XML file can be generated to be integrated with CI tools. Also, as the translation from human syntax to 'machine code' is done automatically, it is different from Cucumber / Gherkins approach, where for each block of human language, there needs to be a matching block of code to be written for execution.
Below are more features of the tool, you can integrate with machine-learning models easily because you can also write Python and R code directly in TagUI scripts to invoke your trained models for inferencing. Or do API calls to those ML services from tech giants.
CLI tool for digital process automation (RPA) -
- automate Chrome visibly or headlessly
- visual automation of websites and desktop
- write in 20+ human languages & JavaScript
- Chrome extension for recording web actions
- R & Python integration for big data / AI / ML
- CLI, REST API, advanced API calls covered
Presenter(s)
Ken Soh - kensoh
Presenter's bio
I'm working as an engineer now at AI Singapore, and maintaining this project as part of my job.
I thought that introducing this to the JavaScript community will be cool, because that is the native scripting language used in writing TagUI scripts (besides using the human language syntax). JavaScript developers would be able to use this right out of the box for their work or personal use cases.
If a speaker bio is needed, something like below can be used or trimmed, otherwise it really doesn't matter. Just hope to put this free and open-source tool into the hands of JavaScript developers.
Ken specializes in using process automation to accelerate & scale digital processes. Prior to joining AI Singapore, Ken founded an open-source company, researching & developing TagUI RPA software out of a 1-year Eastern European sabbatical.
He formerly held technical leadership positions in DBS Technology & Operations, HP Enterprise Services, ADP Dealer Services and A*STAR I2R. Ken is an e-Resident of Estonia, a participant of UN Global Compact and a patron of Mastodon open-source social network.
Duration requested
20 minutes including Q&A? The talk with intro, demos and link to resources, can be done in 15 minutes. I can also go into more details to make it longer depending on how much time there is at the meetup.
Deck URL (if any)
https://prezi.com/p/f5vag20tuth8
Talk Title
Building fast and semantic inputs in VueJS
Talk Description
Forms are commonly notorious for having a poor user experience. When built well however, forms incentivize users to complete them quickly without sacrificing inaccuracies. In this talk, we will explore patterns and strategies to building semantic and robust input elements in VueJS and hopefully learn a thing or two about building things in Vue.
Presenter
Divya Sasidharan
Bio
Divya is a web developer who is passionate about open source and good documentation.
Duration
10-15 minutes. The talk will be a demo of building form input elements using VueJS and will give a high level overview of how Vue works.
Deck URL
TBD
Hello guys! Thanks for the talk submissions, cool stuff. Unfortunately time is a little limited because we have 4 talks scheduled this month so lets work with the time allocations as follows:
- @cheeseprata - 15 mins + 5mins Q&A
- @sophieyang0813 - 10 mins + 3 mins Q&A
- @kensoh - 10 mins + 10 mins Q&A
- @shortdiv - 10 mins + 3 mins Q&A
Let me know if there's any issues! :) I can be reached via the Gitter group.
@kensoh - Cool slides. Based on some experience, I recommend skimming through the reasoning behind this tool and focusing more on code demo (I see a 3mins video) and going more into the technicalities. That would be impactful and helpful. Feels like yours would be one that would benefit from keeping more time for Q&A out of the 20min block.
@cheeseprata Hey, unfortunately we're running 4 talks this month so a 30 minute block would be tough. Let's try to keep the presentation to a 15mins and if you go a little over, that's cool (under 20 please, keeps the audience with you). Let me know if that works. Looking forward!
Thanks @ongmin that sounds like a great idea! I'll do @kensoh - 10 mins + 10 mins Q&A 😃
Title: Frontend Dependency Tracing for HTTP/2 Server Push
Description:
I am working on a tool that automatically generates a resource manifest. I will talk about npm packages and the concepts (i.e. Abstract Syntax Tree, Visitor, and Walk) I learned from building this. As of today, this tool can parse html, css, and js files in my static portfolio directory and write out dependencies. This project is inspired by Server Push Cache Manifest: Eliminating round trip times at the HTTP/2 protocol level.
Presenter
Sophie Yang
Bio
I am a 6 months old baby coder and would love to share my learning experience.