One of my newly found interests during the course of my graduate study is technology entrepreneurship. This didn’t come by accident because I actually worked for a startup software company after I graduated from college. That work experience inspired me to take up my master’s degree because I realized that I was still lacking advanced technical skills.

Thus, as an extracurricular activity, I volunteered for the Philippine Emerging Startups Open (PESO), a local organization which hosts a technology business plan competition. I did only minor tasks but the experience was still worth the effort. In the culmination night for the year 2006, the competition finalists, venture capitalists, the media and supporters of technology entrepreneurship converged. In that event, I was fortunate to have met Peter Valdes, considered to be one of the 10 most inspiring Filipino technology entrepreneurs. I summoned my courage to introduce myself to him. And it was set. I would work for him as a research intern for his company Vinta Systems.

A.I. buzzwords

It was the summer of last year when I started my internship for him. I didn’t have a summer class, thus, I could concentrate on his tasks which I find quite interesting. We seemed to share a common interest in operations research and artificial intelligence, though the latter was a relatively new concept to me. It was during this internship that I learned some A.I. concepts such as neural networks, support vector machines and particle swarm optimization (PSO). I was introduced to the intuitive and fast-converging advantages of heuristic techniques, although they may only lead to suboptimal solutions. I concluded that these serve as good complement to classical optimization techniques, e.g.  when the latter takes too long to converge to an optimal solution.

A little of Python, a little of R

Aside from A.I. concepts, I also got to practice some programming skills. I admit that one of my worst mistakes in college was to neglect my Java programming skill. Thus, in my first job, I ended up doing mostly Excel and MySQL database programming, instead of doing hard-core programming. But then, I eventually realized that it seemed a little more fun just to do prototyping, primarily due to its interactive nature. And it was just the case during my Vinta internship.

My first prototyping task was to code the PSO algorithm in Mathematica. Then I came across Python, seemingly a much better environment for prototyping purposes. What we did in our research experiments is to come up with an idea, and implement that idea in Python. I believe the test cycle is faster compared to creating a similar design in C++ or Java. I also got to learn a little of R, an environment for statistical computing. I find it harder compared to Mathematica and Python, because both programming and statistics knowledge are required. Fortunately, I’m still able to code a few useful scripts such as a comparison of statistical and A.I. techniques on a test data.

Some difficulties

My internship at Vinta Systems is truly a learning experience. I should have known about it during my college years. Still, it’s not without its difficulties.

The A.I. and statistics concepts introduced was relatively new to me. It was quite a challenge to absorb them. If there’s one thing I realized from attempting to understand these concepts, it’s that a good book is indispensable. Luckily, I can just borrow books from the Ateneo Rizal Library and the U.P. Library. Another challenge I had to deal with was working mostly alone on the tasks. Sir Peter was there occasionally to supervise me, but he also had to attend to other responsibilities in the company. Realizing this problem, we tried looking for other researchers, but got little response. I knew our research would be dealt with faster if at least two heads were consistently working on it.

Final words

A few months ago, I decided to stop my research internship for Vinta, as I had to concentrate soon on my studies and my master’s thesis. If I had just more time and someone to work with, I would have continued on our research. Still, this is a type of company I look forward to working for after obtaining my master’s degree. In fact, it’s the kind of firm I’d like to build myself someday.