Success Stories Startup Fintech Company Needs Scalable Site to Support Massive Traffic
The Project
When some of the greatest minds in the Financial industry decided to start a FinTech company, they needed a great technology partner to provide the “Tech” to their “Fin.” They quickly chose Winmill Software.
The company’s goal was to take the mystery out of quantitative investing. They wanted to bring multi-factor and artificial-intelligence models to self-directed investors via an easy to use website and mobile app. Their proprietary algorithms ran on a massive AWS server farm, producing buy/sell recommendations with proven results. And behind the scenes, Winmill Software was making it all happen.
The project was enormously successful. The website won multiple design awards, and the underlying models produced results that quickly garnered the attention of the investing world. After just 18 months, the company was acquired by one of Wall Street’s most prestigious firms, and plans for version 2.0 of the platform are already underway.
Challenges
- The investing algorithms required massive computing power, quickly overwhelming their brand new, heavy duty servers.
- The company did not have the time or money to operate its own IT department.
- The website had to be a home run. It had to be beautiful, engaging and intuitive. It had to be able to present massive amounts of information without being overwhelming.
- The website had to access stock information from a major information provider, but the API’s were too slow. There was an option to use a local version of the database, but it contained in excess of one terabyte of data.
- Site traffic could jump from 50 to 50,000 in a matter of minutes. The site had to be infinitely scalable.
Solution
- Our solution was built entirely on AWS.
- We used massive AWS servers to run the algorithms, taking full advantage of pay-as-go computing. For the 18 hours every day that the algorithms were not running, the servers were idle and the cost was negligible.
- We used multiple indexing and caching strategies to optimize access to the stock database.
- We built a serverless architecture that could rapidly scale. We used Load Impact to run load simulations and identify (and ameliorate) bottlenecks.
- We built application scanning directly into our DevOps pipeline, to ensure that security was incorporated into the application from the ground up.
The Results
The website won rave reviews and multiple awards for its design and user interface.
The platform received high acclaim in the industry, quickly becoming the default site for self-directed quantitative investors.
The site was infinitely scalable; new servers would automatically spin up at predetermined traffic thresholds.
The ultimate mark of success was the acquisition of the company by a major Wall Street investment bank. They are in the process of incorporating the platform into their own offerings, and will soon initiate development of Release 2.0.
