18Fall CS PHD文书求改求批判~

这是我写的PS，然后找了一个工作室帮改了一下，求大家修改意见~

My intense interests in computer science are driven by the passion I have to understand how technology gadgets work. From an early age I have been studying programming and creating software programs. My time is consumed by computer learning and the choice of enrolling in computer science in undergraduate school came naturally to me.

I will receive a bachelor degree in computer science and technology from the University of Electronic Science and Technology next summer. The school ranks among the best in computer science in China and my academic performance has been outstanding. I have not only received several scholarships and competitive awards for my efforts, but also have significant research and internship experience to help make the transition to Ph.D. studies easier. Two of my research papers are in the process of being published.

Last year I joined the Zebra laboratory and contributed to the FlexBM project. My research involved a dual-scheme block management solution for distributed systems that would combine the benefits of EC coding and replication, serving as a supplement to existing Hadoop-like solutions that suffer from either excessive storage overhead or degraded reads. My specific tasks on the project included performing experiments on the modified Hadoop system to evaluate performance of the FlexBM by comparing with previous studies such as HDFS-RAID. The experimental results indicated evident effects of this adaptive block management policy on accelerating task execution, and at the same time the storage overhead was reasonable.

More recently, I carried out a major research project to speculate the I/O behavior of distributed applications in a lightweight way, aiming to alleviate deficiencies of existing I/O profiling frameworks such as shortage of portability and being costly. Based on the empirical observation of the connection between metadata access and data I/O, SwiftIO was found as a meaningful solution from the correlation through extensive experiments. After I initially experienced some setbacks with my approach, with the assistance of my supervisor, ways around the unresolved issues were realized. I used a series of MapReduce applications, did extensive testing, and made refinements in my processes to get to where the experiments were able to prove strong linear correlation between data I/O volume and metadata access metrics on distributed storage systems. Concurrently, a prototype implementation of SwiftIO algorithm verified the feasibility of predicting I/O behavior accurately through metadata access in a lightweight manner with extremely low overhead. My research has been submitted to a high-end academic conference, and could potentially have positive outcomes for I/O profiling technologies.

Apart from my research experiences, I also have involved myself in quite a few internships. Two years ago, I worked for Brithon as a full-stack software developer. My specific tasks were on the online appointment scheduling service (BirchPress) for the development of a CRM module using ReactJS and PHP. Currently I am working part-time on a project called Overseas New Life, and my primary obligation is being in charge of a WeChat bot service that provides auxiliary functions for online merchants to serve their customers. Two particular challenges in this position that I have overcome involve the integration of the bot framework into the server, and the memory leak issue. Since the framework we used was designed for clients, changes were required before applying it to the backend application that holds multiple users simultaneously. I successfully developed the server-side bot platform by applying multithreading and co-routine, enabling the server to handle numerous bots of different users with customized configurations. As well, by limiting the message queue, and reusing instances via Redis-based session, the memory leak problem was mitigated, allowing the server to operate smoothly.

My experiences in research and internship work not only have expanded and enhanced my skills and ability to learn, but they have also let me recognize some of the trends occurring in technology. I realize that we have entered an era of mobile internet and big data, where people become increasingly dependent on connected electronics, while data is being collected for making further changes for our convenience. Moreover, it is my opinion that advances in technology such as artificial intelligence and augmented reality require reliable underlying systems to support them. As applications continue to expand, demands on high performance and robust distributed systems will increase. At the same time, there will be ever more challenges for making optimizations on distributed systems regarding performance and efficiency because of limited budgets of startups, finite resources of energy and the physical limits on transistors.

The Ph.D. program in computer science at University of XXX will not only prepare me for a rapidly changing world but also give a meaningful way for me to make contributions to society. It is my hope to work in the XXXX lab to be involved with the most advanced research on distributed systems. XXXX has an impressive set of accomplishments, and I am intrigued by the research being done on reliability and performance improvements concerning distributed storage. My specific research interests closely align with those of Professor ####, such as the topic of SSD optimizations and crash consistency. My long term career aspirations are to use the knowledge I gain to work for a large internet company or cloud service provider as a researcher and provide solutions for the digital world of our future.

I am confident of excelling and making positive contributions to your department’s research initiatives. Thank you for your consideration of my application.

Wei Su

Prospective PhD Student

1. 第一段其实可以删掉吧，第二段夸学校夸自己的内容也没什么存在的必要。
2. research experience写得蛮好的，很清晰，可以考虑每段后面加一句impact sentence，讲一讲how did this particular project shape your research interest; what did you learn from this experience. 同时你也可以考虑加入一些关于soft skills的内容，因为搞科研并不是仅仅technically strong就可以的，你还要有good communication skills, perseverance, teamwork spirit, skills in life-work balance等等等等。
3. internship的篇幅是不是太长了。
4. future goal/research interest有点太弱了。其实这里是你展示自己作为一个researcher的vision的地方，不仅要说你和哪些教授的方向相关，更重要的是你自己要研究什么问题。下面复制berkeley的SoP指南:

Here you indicate what you would like to study in graduate school in enough detail to convince the faculty that you understand the scope of research in their discipline, and are engaged with current research themes.

a) Indicate the area of your interests. Ideally, pose a question, define a problem, or indicate a theme that you would like to address, and questions that arise from contemporary research. This should be an ample paragraph!

b) Look on the web for information about departments you’re interested in, including professors and their research. Are there professors whose research interests parallel yours? If so, indicate this. Check the specific program; many may require you to name a professor or professors with whom you might work.

c) End your statement in a positive manner, indicating your excitement and readiness for the challenges ahead of you.

nju-wjr 发表于 2017-12-7 21:41
对于第一、第二段以及 intern 的看法（话说这个 intern 和 LZ 的方向好像关系不大？）和前面的大大差不多
...

thank you~加油~

对于第一、第二段以及 intern 的看法（话说这个 intern 和 LZ 的方向好像关系不大？）和前面的大大差不多

“My experiences in research and internship work...” 这段感觉谈的有点泛，感觉能给出一些 LZ 方向的 insight 就好了

同申 CS PhD，方向不同，彼此加油吧

话说咋设置阅读权限啊，这样直接贴出了会不会被中介什么的盗用。。

写的太好了，，感觉不光会被中介盗用，，也会被地里面的同学参考。。。。

Britjeans 发表于 2017-12-7 18:37
写的太好了，，感觉不光会被中介盗用，，也会被地里面的同学参考。。。。

同学参考没有意见啊，就是不想让某些机构的采集器拿去。。。

申phd的SOP的话，应该加上未来想做啥吧。另外一些项目感觉仅仅是陈述，是否应该加一点解释，做了这些项目能对你的phd学习有什么帮助？

andylee 发表于 2017-12-7 18:47
申phd的SOP的话，应该加上未来想做啥吧。另外一些项目感觉仅仅是陈述，是否应该加一点解释，做了这些项目能 ...

嗯…最后两段就是未来想做的吧

gqyz01sw 发表于 2017-12-7 18:49
嗯…最后两段就是未来想做的吧

仔细看了一下，挺好的，加油！

只提一个教授或者单独的一个实验室会不会有风险额

大都会做题家 发表于 2017-12-7 19:05
1. 第一段其实可以删掉吧，第二段夸学校夸自己的内容也没什么存在的必要。
2. research experience写得蛮 ...

谢谢建议~