疯狂工作 II

公司里的印度人 赶deadline还在打游戏 开party 很放松的气氛 我觉得这很有益于我的工作。

不急不躁 困难的东西用最容易的方法弄出来 长时间工作 但是气氛很好 

这真是我拍脑袋想不出来的工作方式啊。

Distinguished Engineer + 疯狂工作

跟着公司的Distinguished Engineer code, 真会爱上coding, 爱了会更加爱。


Vk
=====
0. Rely entirely on logic, not on syntactic similarity.
1. Code slowly (type slowly) and calmly even under deadline, able to understand the code with the whole picture without running it, and change them with confident;
2. If things are not right, do not do a quick fix, fix the right things by thinking of a proper architecture;
3. Focus on how to produce nicely knitted code that blends well with existing codes;
4. When undo etc., look at each places and figure out whether they are correct;
(codes is like an article, look at it without running)
(figure out the best flow for handling things, with good naming convention)
5. Very patient, no rushing for finishing.
6. Don't treat code improvement as a deadline thing, it is a continual process, treat it as a hobby thing
7. Have a clear workflow in mind, can immediately point out which workflow is redundant, which can be simplified

Code Advice
1. reduce parameters (copy utterances, new_mentions)
2. de-nesting
3. de-memory for long code
(no visited variable, if not...return directly rather than go last)
4. no bidirectional code
5. don't scare of correctness, just refactor and use unit test for your advantage
6. don't call redundant things
e.g. when you have candidates already, don't call a function where first step is get_candidates
7. annotate the output
a)Yes
if not ..
return
b)No
if ...:
long list

8. make a new list rather than delete element in existing list
9. a. function 1,2,3 return
Can just move it outside and detect for that instance
[each function should focus on one task, not too many things;
and don't put specific things too deep;
think of why you feel easily lose track of what is done in the current code]

Additional
1. use extract method and extract variable to comment the complex chunk of code and complex chunk of expression.
2.

这几天特别专心的工作，因为一个deadline 接着一个deadline.
学了很多东西，因为一群印度人从印度过来。
他们真是一群乐观的民族，deadline的时候也没见他们疯狂工作，就是打游戏什么的；然后晚上再工作。工作中有娱乐，娱乐中有工作。

自己方面，顺利完成目标有种成就感，但是这有赖于旁边的三位印度人的帮助。
开始的时候完成了code很开心，然后那个distinguished enginner要求写unit test才能check in, 我还以为可以慢慢写，我也没写过test case，于是又专心的码着，我觉得这种逼迫对学习很有效，几个小时内从新手变成中级用户。完全没有感觉时间的过去。deadline真是一把好刀。

从本来的不情愿加班，到加班加到很有热情，中间的成就感很重要。我觉得在工业界需要这种热情来带动，而不是单纯的工钱和时间的对等，这样才能做的积极快乐，学的更多，这可比金钱价值高多了。这怎么看都是一种良性循环。

另外发现很多是后最不想干的事情都是最重要的事情。先干了再说。

享受工业界

我觉得有几点很不习惯
- 常常开会 （训练开会能力，训练开会的同时处理事情）
- 常常变动 （训练领场表现）
- 常常engineer (训练engineer能力)


虽然觉得没学到什么 但是大多数的学习都是隐性的。
除了NLP方面，更多的是其他工业界需要到的根本技能。

要享受工业界，也必须享受这些技能的学习和运用。
不要回避，去面对，转化，任运。

用时

感觉今天累点，但是accomplish的东西还挺多的
主要 -
１．把要做的东西列出来，然后把他分到每天的schedule里去
２．早点起床，开始工作

主要发现
１．你也不想马上做的东西，越要赶紧完成。因为这些不想做的东西，往往是最重要，需要去思考的东西。

Meeting

XX are you there?


SJN
What is the failing? (When say the problem is due to failing)
What are the next tasks? (when say there is some next task)
I want to wrap up, there are other pressing tasks there has a higher priority
I hope you understand, if there are other tasks of higher priority, and such tasks there is no relevant to the launch, I don't see a point to spend time on it.
I have no further question, and I know fairly well on Aqumar..Tian Update.
Ok sure, and go ahead.
xx you want to go next

Ranu
To get data from ANY party if required (classifier code - Vandan; training data - Hari)
To understand the foundation, where do you do the coding (NLU PMP)
To get help from ANY party (training data - Alexandra)
To get things that remove the stress of deadline first, and to be review by other people (get correction v1.0 done first and get review)

1. what need to get, and
2. who can give help
3. where to start from

4. pin point the exact things we want to handle (define what Hari use cases to be handled)
5. get next things ready (get Sindhu ready for the test cases)

6. Weekly plan with concrete output
7. Tabular analyis of all possible cases

Other
I will concentrate primarily on ...
Yes, SJN, but there is an issue on...

追求

发现人的追求是无止境的，有了一样还想着另一样，于是乎就无止境的努力着。
但是想想这何尝不好，中间有快乐痛苦，但是这种正向的思维努力，是推动积极正面的重大因素之一。

反省一下自己的时间利用:

Reflection
Waste: 感觉自己最近常常花时间看新闻，看赚的利润，睡觉等等。这些都是0 return的extra work。

Interview Prep: 回想一下过去面试的经验，一般会问些什么问题，怎么去回答+impress - Big Vision, the problem, how you would address the problem, benefit。都是套路，但是我们要从这个层面下手。

Side Work: 基本处于一种最重要的东西还没弄出来的感觉。最重要的是collect contact + create value and make it free。

自我修行: 其实更多的是对自我行为反省。觉得自己有点懒，心理浮动有点大，同理心有待加强。

Work: 都搞一些superficial stuff。Tree-based CRF 和 anaphora resolution都没有时间看看。

Action Items
Waste: 砍了。

Interview Prep:  
Data science - 从广的层面，use cases 来下手，不用太Theory。
Algorithm - 要练练 (天天一小时)
SQL - 要看看。
录个Video 以后可以看看。

Side work: 搞个Email collection, 搞多点free stuff.

自我修行: 天天反省。

Work: 专心，多花点时间， 少吃东西，少看别的东西。

The Crest

上个周六终于把房子搞定了。
本来去看The Cape, 后来中介发现我喜欢view，多过于户型，于是他就把我带的The Crest.
当天view 最好的三房开了四间来卖，剩下一件。
于是周六晚上一阵纠结，然后当天就决定买了。
Condo lock,周日给支票。
幸好defer payment, 压力小点。

让我想起四年前的Coco palms购买，剩下两间，拿了一间，最后剩下的一间隔天卖完。
感觉我们都是误打误撞到抢房的阵地里去随风起舞。