微软hackathon
by Ajay NS
由Ajay NS
Hackathon报告：30小时内您可以编写什么代码？ 非常多！ (Hackathon Report: What can you code in 30 hours? Quite a lot!)
What can you build in 30 hours straight? As a group of second year college students with a growing portfolio of work, my team and I wanted to find out. So we signed up to a hackathon.
您连续30小时可以建立什么？ 作为一组二年级大学生，他们的工作量越来越大，我和我的团队想找到答案。 因此，我们报名参加了黑客马拉松。
It was a Financial Technology (or ‘Fintech’) hackathon organized by DCB Bank in the city of Mumbai. Although we were clueless about Fintech, we wanted to give it a try, in the hope of coming with an idea that solves a general problem.
这是金融技术(或“金融科技”) DCB Bank在孟买举办的黑客马拉松。 尽管我们对金融科技一无所知，但我们还是想尝试一下，以期提出解决一般问题的想法。
The event was hosted in the beautiful and cozy co-working space by 91 Springboard. This was an environment I hadn’t been in before.
该活动由91 Springboard在美丽而舒适的合作空间中举办。 这是我以前从未去过的环境。
Basically, it rents out office space for startups, freelancers and such who don’t need a huge office, but just a workspace for their team members to collaborate. 91 Springboard takes care of providing fun colorful and snug workspaces, fast internet and unlimited coffee, while you work hassle-free. It was really a pleasure crashing here for the weekend.
基本上，它为初创公司，自由职业者等不需要大型办公室的人出租办公室空间，而只是一个供团队成员进行协作的工作区。 91 Springboard致力于为您提供丰富多彩，舒适的工作空间，快速的互联网连接和无限的咖啡 ，同时让您轻松工作。 周末在这里撞车真是一种荣幸。
想法 (The Idea)
To put it simply, we wanted to build a specialized chatbot.
简而言之，我们想要构建一个专门的聊天机器人。
Why?
为什么？
It’s predicted by 2020, a huge percent of business communication will be done through chatbots.
据预测，到2020年，很大一部分业务通信将通过聊天机器人完成。
User experience (‘UX’) is rapidly changing. With messaging apps gaining popularity, users prefer to access everything under one app. They are also in favor of one-on-one communication.
用户体验('UX')Swift变化。 随着消息传递应用程序的普及，用户更喜欢访问一个应用程序下的所有内容。 他们也赞成一对一的交流。
Gone are the days of emails, it’s all about real-time chat now. With today’s technology, it’s possible to create a chat-bot which can learn by itself and automate most of the tasks, allowing mass communication on an individual level.
电子邮件的日子已经一去不复返了，这一切都与实时聊天有关。 借助当今的技术，可以创建一个聊天机器人，该机器人可以自行学习并自动执行大多数任务，从而可以在个人层面上进行大量交流。
See here for a list of some of the best uses of chatbots right now.
请参阅此处以获取一些 聊天机器人的最佳用途。
我们到底建造了什么？ (What exactly did we build?)
As we were at a Fintech hackathon, we thought about a bot that carries out all the functions that can be carried out by the bank’s app and more — through chat.
在我们参加金融科技黑客马拉松时，我们想到了一个机器人，该机器人可以执行银行应用程序及其他所有功能的所有功能-通过聊天。
Once logged in, you can ask the bot for your balance, your last few transactions, and also to carry out actions such as fund transfers. It leverages the required APIs provided by the bank for the same purposes.
登录后，您可以向漫游器询问您的余额，最近的几笔交易，还可以执行诸如资金转账之类的操作。 出于相同的目的，它利用了银行提供的必需API。
Above is a screenshot of some of our UX inspiration. This UX had a combination of natural language and multiple options, which were directly provided as links and buttons. The flow of answering users queries, suggesting the next steps and also allowing actions to be carried out, was planned.
上面是一些UX灵感的屏幕截图。 该UX结合了自然语言和多个选项，这些选项直接作为链接和按钮提供。 已计划了回答用户查询，建议下一步并允许执行操作的流程。
该项目 (The Project)
The tech stack we chose was Ruby on Rails. We picked this because my teammates were very familiar with it, and I was focusing on wrapping up APIs and UI primarily. But this stack was totally new to me, as I’ve always worked on JavaScript or Python stacks. I wrote about my learning journey here.
我们选择的技术栈是Ruby on Rails。 我们之所以选择它，是因为我的队友对此非常熟悉，并且我主要专注于包装API和UI。 但是这个堆栈对我来说是全新的，因为我一直在使用JavaScript或Python堆栈。 我在这里写下了我的学习历程。
It uses a basic PostgreSQL database to store user messages, and uses ActionCable for live data-streaming as required. The whole site (along with the chat as a widget) is built over the Materialize framework.
它使用基本的PostgreSQL数据库存储用户消息，并根据需要使用ActionCable进行实时数据流传输。 整个网站(以及作为小部件的聊天)都基于Materialize框架构建。
One of the key features we planned to use was a customized AI, rather than going for something like IBM Watson or api.ai. This was wrapped into an API using Flask.
我们计划使用的关键功能之一是定制的AI，而不是像IBM Watson或api.ai这样的东西。 使用Flask将其包装到API中。
Initially, when the bot has little to no training data, human assistance is required to give responses and also classify the intent of each user query for the bot to learn. Once it accumulates data, it can automate the whole process, giving appropriate responses.
最初，当机器人几乎没有训练数据时，需要人工协助来做出响应，并对每个用户查询机器人的意图进行分类。 一旦积累了数据，它就可以使整个过程自动化，并给出适当的响应。
A few of the scripts used for the AI are available here.
黑客马拉松 (The Hackathon)
We managed four hours of sleep over a span of more than 30 hours, while pushing over 50,000 lines of code to production.
我们在30多个小时的时间内管理了四个小时的睡眠，同时将50,000行以上的代码推向了生产环境。
But it wasn’t as bad as it sounds! The place itself has this motivating vibe, and there are mentors helping you out. There were breaks in between to chill with other teams and getting to know other kick-ass developers.
但这并没有听起来那么糟糕！ 这个地方本身就具有这种令人振奋的氛围，并且有导师帮助您。 在这之间有一些休息时间，可以让其他团队放松，并认识其他踢屁股开发人员。
While my team mates focused on the Machine Learning part and the actual chat application on Rails, I worked on the UI for the service. This included an admin panel as well as the chat widget. Quite a lot of time was spent wrapping up the AI code into RESTful APIs and fixing bugs in the main code as well.
当我的团队成员专注于机器学习部分和Rails上的实际聊天应用程序时，我在该服务的UI上工作。 其中包括管理面板和聊天窗口小部件。 花费了大量时间将AI代码包装到RESTful API中，并修复了主要代码中的错误。
In the end, what we had was a basic alpha version of the app we set out to build.
最后，我们拥有的是我们打算构建的应用程序的基本Alpha版本。
体验 (The Experience)
For a start, this was the most productive 30 hours of my life.
首先，这是我一生中效率最高的30个小时。
With constant bursts of motivation (and caffeine) driving me, there was no time for sitting around or lazing. Loads of fun and talented people to interact with as well, because I was one of the youngest there.
在不断涌动的动力(和咖啡因)的驱使下，我没有时间闲逛或闲逛。 也有很多有趣的和有才华的人与人交流，因为我是那里最年轻的人之一。
A few of the things I learned here included:
我在这里学到的一些东西包括：
- Decoupled application architecture 解耦的应用程序架构
- Building an app from scratch and then pushing into deployment in the shortest time possible 从头开始构建应用，然后在最短的时间内进行部署
- Live code collaboration 实时代码协作
- and of course, a little bit of Fintech! 当然还有一点金融科技！
Check out the project we worked on here: chaturbots.com. This is offered as a service now, so contact us for getting a custom bot built.
在这里查看我们从事的项目： chaturbots.com 。 现在提供此服务，请与我们联系以构建自定义机器人。
Hope you enjoyed this article and found it to be a good read! You can check out all my projects on Github and reach out to me on Twitter!
希望您喜欢这篇文章，并希望它能对您有所帮助！ 您可以在Github上查看我所有的项目，并在Twitte r上与我联系 ！
微软hackathon
1098
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
