Profile cover photo
Profile photo
Share new ideas on cloud, big data and other hot tech areas
Share new ideas on cloud, big data and other hot tech areas


Google HFT? Amazon Trading Service? MS Financial API?

I know a few traders working at home by themselves, and believe that there are a large number of such traders. The key items for them to be successful are market data, security data, programming API and reliable server. There are solutions on the market, but most of them are quite expensive, not reliable and have much worse performance than the ones used by banks and trading firms. This put them at big disadvantage. The situation also applies to all retail customers, though not all of them trade in a rational manner.

In the past couple of decades, Island and other ECNs disrupted the traditional Nasdaq and NYSE markets, and and other online discount brokers gave retail customers direct access to exchanges and brought down trading cost. With the maturity of cloud business and emergence of Big data services, it is time to bring up the next revolution for the trading business: Leveling the trading field by offering all retail customers with fast and clean market data, powerful DMA API, reliable hosting service, and strong research capabilities.

Currently, Google, Amazon, Microsoft, IBM and Apple all have the ability to provide this service. One or more of them  should use their large customer base to negotiate with exchanges/brokers. They can then implement trading platforms on top of their cloud/big data platforms, and offer the trading services to customers using the Gmail-like free approach or AppEngine-like low-rate pay-as-you-go-approach.  This will be a very profitable new business, and at the same time, a good public service to the retail financial customers.   
Add a comment...

Web Persistence

Internet is ubiquitous and we search, read and bookmark everyday. Two issues are raised in this process: 1. Our bookmarks can quickly become junk yard; 2. We worry about that the bookmarked pages/documents may disappear some time later, so we have to save them to local disk. This brings another issue: We save files in multiple computers, and can not easily put them into a single place and access them everywhere.

A cloud/local combined data storage and organization service can solve the issues. Not like photo sharing and other cloud services that store personal files, this service should focus on storing and organizing personal selected pages/files from the Internet. The narrowly defined scope brings the following benefits:
1. Corporate friendly: Since the service does not allow users to send files/documents out, companies do not need worry about data security issue. They will be more willing to allow the usage of this valuable research service in work computers.
2. Green: Since the data is from the public Internet, we can easily identify whether multiple users are storing the same content and only store a single copy on the cloud server. We will certainly ensure each user to have privacy on his/her own meta data.
3. Pages recommendation: Users organize and tag pages/documents for themselves. The cloud service can summarize the information and provide recommendation to other users whenever they search or put new pages.
4. Tag recommendation: Tree structure like the current browser bookmark system is a good way to organize data. But users will also want to assign tags to documents/pages to better organize and search them. The cloud service should smartly recommend tags to users.
The system should run seamlessly on both mobile and desktop environments. It should work as following:
1. A user open this WebPersistence service in a browser tab or app
2. He browses the web in other tabs
3. When he is ready to save a page/doc, he drags the URL to the WebPersistence  tab/app.
4. The tab/app accepts the URL and shows the user's data tree, tag fields, and tag recommendation. The user can organize pages easily.
5. The tab/app also recommends highly related docs/pages to the user.
6. The pages/docs will be stored on the cloud and cached locally. A user can universally access the files while getting them instantly in most case.

To implement this system, some tools and frameworks are  needed/recommended:
1. Intel XDK for front end development. XDK allows you to create beautiful Webpage/App for multiple devices easily. 
2. Django for back end. This allows you to get high performance, scalability, while getting Python's rich scientific libraries. You can also easily host the system in Google AppEngine if preferred.
3. Google BigQuery for fast and large scale data query. The cost may be an issue, though.

Add a comment...


This project is to develop a software-hardware solution that allows users to take picture of themselves without the help of a third party. 

1. The product is a replacement of the self-photoing functionality of the current cameras. It resolves the aiming difficulty and allow users to adjust angle and zooming on-the-fly.
2. When people want to take pictures of their whole family/party outside, they do not need to find others to take their pictures.

We will use one smartphone as the camera and another smartphone to view the video feed of the camera and control it. The camera smartphone will be mounted on a small robot. The robot is controlled by the second smartphone and can adjust the hight, the left-right angle, the back-forward tilt and the horizontal level of the camera. 

The first smartphone camera may also be an android-powered camera instead of a smart phone. The android-powered camera can provide zooming and better picture quality.

Project Outline
The project will be implemented in multiple steps:
• Develop android and ios apps that allow one smartphone to view the video feed of another smartphone and control the other phone to take picture.
• Develop a robot using EV3 to hold and move a smartphone.
• Enhance the android and ios apps to control the robot
• Add support for android powered camera
• Industrialize the solution: Use the smartphone to replace EV3 block, and use small motors to create a portable/sturdy solution.
Add a comment...

OCR-Pen note taking and cloud information storage

I read books on commute train to NYC every day, and mostly the books borrowed from local libraries. From time to time I felt I have to take notes on the useful information such as names, strategies and numbers that were mentioned in the books. For my laziness or inconvenience on the train, I seldom wrote anything down, and I only got fuzzy memories of the books eventually.

A software and hardware combination solution will solve the problem nicely. First, we use an OCR-Pen to capture the important sentences in the book, and the pen communicates with your mobile phone using Bluetooth. An app on the mobile phone will publish the information to a cloud server. The app and the cloud service handle OCR and data organization functions.
Add a comment...

Post has attachment
Welcome to Ideas4StartUp! I am a quants trader working in New York city. I used to work in tech industry and founded a web/video conferencing company. Many ideas ran into my mind lately but I do not seem to have enough time to deliver them. I plan to publish the ideas, initial designs and/or prototype implementations here.  Hope this will help you on your venture.
Add a comment...
Wait while more posts are being loaded