I make computers do my bidding.
I enhance lives using technology,
liberating people from monotonous tasks.
When I first started working in an office, I noticed something perturbing. Easily 70-80% of the time spent was doing incredibly mundane tasks: copying and pasting between spreadsheets, manually generating reports, and other easily defined, yet repeatable tasks. That realization led me to start developing automatons to automate these monotonous tasks and give people the time freedom to do more interesting and creative things.
The bulk of my work in this space focuses on introducing automation and artificial intelligence into information-gathering, and decision-making processes in order to create more consistently applied solutions to known problems.
Increasing efficiency 70-fold.
Before the introduction of the bid management tools, bids on advertising partners were updated manually--a slow and often error-prone policy. The new system now allows bids to be updated in almost real-time in response to incoming attribution data from our advertising partners.
Finding over and under-performing segments in my sleep.
Old Method:
A yield analyst would...
Uber contracted my firm to generate new first-rides for them. We applied a ready-fire-aim strategy to test all possible placements, scale up the ones that were working, and scale down the ones that were not.
The advertising images that were allowed were very restricted, which left us to focus on the placements where the advertisements were shown. Every morning, the previous day's stats would be evaluated by a yield manager who advised the media buying team what bids to adjust.
With the new automation platform in place, we were able to identify micro-segments that generated desirable, higher-quality leads for Uber.
In the run up to Super Bowl 49, Draft Kings contracted us to manage a substantial advertising budget. So we focused again on a ready-fire-aim strategy. The biggest advantage to this strategy are the new patterns and corresponding explanations that I could never have guessed on my own. These sorts of relationships between peoples demonstrated interests--the websites they visit--and the amount of money they deposited with Draft Kings.
These relationships werre easy to extrapolate to other similar advertising placements which in turn led to even more deposits making this advertising campaign wildly successful.
One new technique that I developed was applying a retargeting-style technique to not only follow individuals through the internet, but have the ads shown to them tell a story from browser-session to browser-session. That story-telling further increased the amount of money deposited to the point where we had $5 in deposits for every $1 in ad-spend.
The fabled maker of Candy Crush, Bubble Witch, and other popular games offered ad-budget to get more users into their gaming ecosystem. We were not allowed to make new ads, we could only focus on finding the highest performing placements.
What was most interesting was how popular King games were with women, and how strongly the followup micropayments data correlated with women's playing habits. With that knowledge in hand, it was a simple matter of using some of the bid-watching data to see how women traveled around the internet and what websites they visited and in which order. That way, we could maximize the number of times a prospective user was presented the opportunity to install the app.
What I find most fascinating about the financial markets are the legions of people who rely almost exclusively on charts to divine the future. In an effort to do something similar, I took it upon myself to learn how to write trading algorithms.
A simple pattern-finding proof of concept
The pattern finding algorithm takes the most recent n-bars, and calculates a moving average for each bar. It then takes the change in the moving average and compares the shape of the seed moving average with the calculated moving average for all bars in the previous history.
The most closely linearly correlated sets of moving average are then used to identify times in the past when that bar pattern happened. When the top correlated matches are found, the algorithm captures what happens after the historical matches. These action-reaction pairs are then projected into the future to determine in what direction the price moved. Whenever the projection has a strong bias in one direction versus another, the algorithm places a trade to benefit from the projection.
Surprisingly, the plan worked fairly well at predicting future movement about 53% of the time. Not enough to really be counted as a strong edge, but definitely a move in the right direction. The trade opportunities are fairly strong, but the system needs to be improved to understand when to exit a trade.
My favorite hobby: Flying at the front of planes.
I afford it with various frequent flier programs to book incredible travel experiences--at the front of the world's premier long-haul carriers--for as much as 95% off of the ticketed price.
I've flown Emirates First Class twice
I've now flown Cathay Pacific's excellent First Class product four times.
I've been on Lufthansa'a FRA-JFK flight LH404.
This one was an award redemption using 70,000 Aeroplan miles. A very worthwhile experience--especially the First Class Terminal in Frankfurt.