IT討論區(77):經濟一差,西工處處

舊post
(1): https://lihkg.na.cx/thread/4990/page/1
(2): https://lihkg.na.cx/thread/69444/page/1
(3): https://lihkg.na.cx/thread/89460/page/1
(4): https://lihkg.com/thread/116590/page/1
(5): https://lihkg.com/thread/130061/page/1
(6): https://lihkg.com/thread/148992/page/1
(7): https://lihkg.com/thread/170835/page/1
(8): https://lihkg.com/thread/208042/page/1
(9): https://lihkg.com/thread/233717/page/1
(10): https://lihkg.com/thread/250870/page/1
(11): https://lihkg.com/thread/269901/page/1
(12): https://lihkg.com/thread/287135/page/1
(13): https://lihkg.com/thread/303147/page/1
(14): https://lihkg.com/thread/323813/page/1
(15): https://lihkg.com/thread/346843/page/1
(16): https://lihkg.com/thread/385374/page/1
(17): https://lihkg.com/thread/407488/page/1
(18): https://lihkg.com/thread/432444/page/1
(19): https://lihkg.com/thread/456486/page/1
(20): https://lihkg.com/thread/471984/page/1
(21): https://lihkg.com/thread/503507/page/1
(22): https://lihkg.com/thread/547628/page/1
(23): https://lihkg.com/thread/576562/page/1
(24): https://lihkg.com/thread/602397/page/1
(25): https://lihkg.com/thread/627728/page/1
(26): https://lihkg.com/thread/651466/page/1
(27): https://lihkg.com/thread/673484/page/1
(28): https://lihkg.com/thread/690968/page/1
(29): https://lihkg.com/thread/701812/page/1
(30): https://lihkg.com/thread/718050/page/1
(31): https://lihkg.com/thread/729926/page/1
(32): https://lihkg.com/thread/739411/page/1
(33): https://lihkg.com/thread/746333/page/1
(34): https://lihkg.com/thread/754096/page/1
(35): https://lihkg.com/thread/762089/page/1
(36): https://lihkg.com/thread/769677/page/1
(37): https://lihkg.com/thread/776679/page/1
(38): https://lihkg.com/thread/784699/page/1
(39): https://lihkg.com/thread/791805/page/1
(40): https://lihkg.com/thread/800016/page/1
(41): https://lihkg.com/thread/807460/page/1
(42): https://lihkg.com/thread/816001/page/1
(43): https://lihkg.com/thread/828694/page/1
(44): https://lihkg.com/thread/837713/page/1
(45): https://lihkg.com/thread/855218/page/1
(46): https://lihkg.com/thread/871851/page/1
(47): https://lihkg.com/thread/895032/page/1
(48): https://lihkg.com/thread/906329/page/1
(49): https://lihkg.com/thread/921824/page/1
(50): https://lihkg.com/thread/933250/page/1
(51): https://lihkg.com/thread/941628/page/1
(52): https://lihkg.com/thread/954097/page/1
(53): https://lihkg.com/thread/966642/page/1
(54): https://lihkg.com/thread/975683/page/1
(55): https://lihkg.com/thread/986392/page/1
(56): https://lihkg.com/thread/1001177/page/1
(57): https://lihkg.com/thread/1018974/page/1
(58): https://lihkg.com/thread/1040860/page/1
(59): https://lihkg.com/thread/1055888/page/1
(60): https://lihkg.com/thread/1081932/page/1
(61): https://lihkg.com/thread/1101303/page/1
(62): https://lihkg.com/thread/1122848/page/1
(63): https://lihkg.com/thread/1144684/page/1
(64): https://lihkg.com/thread/1173865/page/1
(65): https://lihkg.com/thread/1221713/page/1
(66): https://lihkg.com/thread/1279229/page/1
(67): https://lihkg.com/thread/1311338/page/1
(68): https://lihkg.com/thread/1364346/page/1
(69):https://lihkg.com/thread/1494024/page/1
(70):https://lihkg.com/thread/1657743/page/1
(71):https://lihkg.com/thread/1698184/page/1
(72):https://lihkg.com/thread/1791753/page/1
(73):https://lihkg.com/thread/1791761/page/1
(74):https://lihkg.com/thread/1845165/page/1
(75):https://lihkg.com/thread/1888137/page/1
(76):https://lihkg.com/thread/1904680/page/1
P仔 頂爛市參考價 ：
乜都唔識 + HD： 16k - 18k
識少少 + HD： 17k-19k
u grad + 白紙： 18k~25k
P仔 頂爛市參考價 ：
乜都唔識 + HD： 16k - 18k
識少少 + HD： 17k-19k
u grad + 白紙： 18k~25k

https://medium.com/@p5d12000/工程師應該放心大膽地創造技術負債-a8022d85810
身為軟體工程師，你應該放心大膽地創造技術負債。這麼做唯一的風險，就只有在你換工作的時候，也會接手一大筆前人留下來的技術負債。不過，這種事情反正也早就已經發生了。

話說 其實我覺得最近啲po吹奏IT吹奏得好唔正常 中家點睇

水人入行

唔開Post 頂你

睇返上個po#14

但我都懶得開

寫個Bot搶1001都好喎

Selenium啦 快寫完又易擴充

D人仲好開心30%increment轉工，過到試用期先再講。

酸咩

等你開Repo

升LV 加30%唔過份

諗到樣野 不過好似係常識
1. 拎晒user同po既dataset先
2. 基於user interaction做一個粗clustering (user interaction=quote)
3. 基於呢個user cluster去project落po space 基於常見term去regime cluster
4. 基於refine左既cluster set 為每個user generate一條binary vector

Usage
1. 做hamming space query去一下子select一大堆有共同特徵既user
2. 係唔同既data source到拎cluster set 然後比對唔同data source既user既binary vector 去做配對
3. 發現correlated既cluster 為user提供佢可能感興趣既領域資訊

咁複雜，似NLP+Tagging咋喎

手巴做好粗糙既substring分詞架咋 手巴唔識NLP架

但廣東話 substring 唔準

你會發現做少少trick之後 2~5既substring+1隻code狗陣唔中maintain一個dictionary 已經夠做
可以用一隻廉價狗解決既問題 無必要用十隻高級狗去research幾年
80% solution比起100%solution 係更平 更快 更安全

Semi Auto

lm

明明個市咁差
點解d agent 狗仲日日call