Skip to main content

Columbia MSEE

Program Overview

Cooking at home + eating out (70% / 30%), the total cost for one and a half years is around 800K RMB (not a particularly frugal lifestyle). Claims on Xiaohongshu about spending 1M RMB per year in New York are pure nonsense and clearly not based on any real research. Someone who can spend 1M RMB per year in New York could just as easily spend the same at UCSD. New York is a great place with rich cultural and entertainment options.

Course Selection

You can take many CS courses, making it suitable for career change to CS.

Three semesters, 10 courses total, at least 5 EE courses, at least 5 courses numbered 6000+. In theory, the rest can all be CS courses. It can be more flexible than CE -- CE has required hardcore courses, while EE allows you to take purely CS courses.

  1. COMS4111 Intro to Databases. Waral dи, Donald's course, covering SQL for relational databases through indexing, plus MongoDB and Neo4j for non-relational databases. About 5 assignments, all coding in Jupyter notebooks, practicing with Python's SQLAlchemy or connecting to MongoDB and Neo4j APIs for NoSQL exercises. Take-home midterm and final exams, very easy to get an A.

  2. COMS6156 Topic in Software Development: Cloud Computing-baidu 1point3acres Also Donald's course, covering software development processes, AWS components, and microservice concepts. Lectures are all conceptual, no homework, no exams -- just one large 5-7 person group project. The project requires building a microservice web app, all deployed on AWS. Main requirements include: RDS, Lambda, EC2, Elastic Beanstalk, Cloud API, S3, CloudFront, etc. If you take it seriously, you can learn a lot, especially about AWS and the serverless trend. Meeting basic requirements gets you an A.

  3. EECS6893 Big Data Analytics Ching-Yung Lin's course -- he's also the CEO of an AI company. Listening to a CEO teach means lots of concepts and few details, so you need to learn on your own. This course covers a wide range: Hadoop, Spark, Machine Learning, Air Flow, d3.js -- each topic is covered and has programming assignments. Very time-consuming, with a final project that also takes a lot of effort. About half the class gets an A.

  4. ELEN6770 Cloud Computing & Network Virtualization This course is taught by two instructors (front and back halves), both from Bell Labs. Relatively easy. Two assignments, midterm and final exams. Grading is about half A (A-) and half B (B+).

  5. CSOR4231 Analysis of Algorithms. Professor Eleni is great, course quality is high, and she hand-writes calculations during lectures. This course focuses on correctness proofs and complexity analysis of algorithms -- it's math-heavy with no programming. Assignments + midterm and final exams.

  6. COMS4705 Natural Language Processing Daniel's course, content is relatively basic, well-organized with detailed slides. Assignments + midterm and final exams.

  7. ELEN6889 Large-Scale Stream Processing Taught by an Indian-American professor with only a slight Indian accent -- well-received. The professor lectures very well, covering stream processing applications such as Spark and Apache Beam.

  8. ELEN6883 Intro to Blockchain Taught by Chong Li, who also offers course 6885 in the fall -- both are easy-A courses. Topics include Bitcoin, NFTs, and other broad blockchain-related subjects.

Admission Threshold

The admission threshold is friendly -- there are data points of 985 university students with 3.4 GPA being admitted, but that doesn't mean this program is bad. Columbia's title is enough to get interviews at any company. There are also students from Peking University CS who come here. Whether you go back to China or look for jobs in the US, no one has ever been looked down upon because of Columbia -- everyone thinks it's impressive. Nobody cares about the admission threshold; only applicants during application season care about it. (The world is vast; the capable shall thrive. Ability matters more than credentials.)

  1. HKUST EE, First Class Honours
  2. Shanghai University EE undergrad, GPA 86
  3. UESTC EE undergrad, GPA 89
  4. Xidian University CS undergrad, GPA 3.84, later successfully transferred to PhD within Columbia EE
  5. Utah EE + Math undergrad, GPA 3.95

Job Outcomes

Has Meta intern and Amazon intern placements. Getting past the resume screen is easy; the rest is entirely up to the individual. Bloomberg also recruits from Columbia -- Columbia is rated as a high-ROI school in Bloomberg's internal system, so they specifically come to recruit and interview (recruited a large batch in 2022).

  1. SUSTech EE undergrad, landed hardware new grad in Texas
  2. Shanghai University EE undergrad, had small company internship in China, landed small US company product manager