Daniel de Souza Carvalho - Partner at A2F Solutions
When I was young I saw some older friends going to technical high school of electronics or computing, and I usually saw them wiring components and building up some circuits as homework, and they gave me some advice on how that stuff worked, and I got really interested in it. But I was interested in computer programming too, and so I decided to move to computing, and learn electronics informally through books and magazines. I was in doubt about going to the Electric Engineering or Computer Science areas, because I like both subjects. For my luck, the Computer Engineering course arrived in Brazil the year I was ready to start in college, and then I got it (at Braz Cubas University), because it covers both electronics and computer science subjects. Latter, before some experience in the IT industry, I had MSc in Electrical Engineering (at Mackenzie University) with major in AI and adaptive systems.
For data center diagnostics, I use the IR Thermal Camera, that shows the difference of heat from servers in racks. It is very useful to solve hot spot issues and get a better HVAC and power efficiency from data centers with just some adjustments.
For computer science research, I use a personal HPC. That is a PC with 8 cores CPU and 128 cores GPU (NVIDIA CUDA), 12GB of memory.
For enterprise application I like to work with Oracle JDeveloper with ADF, It is a good framework where the repetitive programming tasks are automated and we can very focus on the system logic instead of being coding around.
For electronics I like to deal with Arduino; I´ve worked on some projects at home as a hobby. And I use a tool kit from Radio Shack with cutters, soldering iron, protoboard, multimeter, third-hand tool and others.
I really enjoy the recent grows of DIY culture; it is amazing the kind of electronics projects we can find out at the WEB; the components are cheap and the results are really cool. Using microcontrollers nowadays is really easy.
The engineering life is made for solving big puzzles on a daily basis. The worst one I can remember was to be on call for 32 hours at IBM solving an enterprise application integration issue with people from India, Brazil and USA together in a production environment. It was a very huge telecom system with many interfaces and components sending XML messages to one another. We found the issue and did a bug fix, and all the problematic XML messages got processed. Enterprise systems are very hard to integrate and maintain.
I like reading books and magazines. My bookshelf covers an extensive list of technical stuff; some I like to revisit include:
Moreover, I like reading Wired, Make and 2600 magazines. I keep a list at Amazon.
I like to use biologically inspired algorithms, such as neural networks, evolutionary computation, DNA, ant colony and others. I model problems as computable data, and do some tinkering about then. I like to process information by different ways to see how it behaves, taking advantage of computational power and tools available. Some surprisingly results come up. Doing some non conventional computation we can find interesting ways to solve problems. I publish some stuff as demo programs. We can use computers to do engineering design and not just work as CAD/CAM tools.
I have many good projects to talk about. Some different stuff from the business projects, I can recall is about the NKS Summer School in 2007 and 2011. It is three weeks ‘hackathon’, where we study and do research in computer science, specifically computation with simple rules, such as cellular automata, Turing machine, graphs and networks.
A real world system can be modeled as simple rule and studied from a computer science point of view. Simple rules are interesting models to study complex systems. As a practical example, it is possible to find programs that can come up with the circuit design, instead of designing the circuit in the traditional form by an engineer.
In 2007, my project aimed at different substitution systems and in 2010 at cellular automata classification with morphological imaging processing techniques. This is the kind of theoretical computer science projects I like to research, and use the personal HPC for them.
From 2005 to 2010, I worked for “Brasil Telecom”- ”Oi” at the designing engineering department with a team of 12 engineers. Ten of them was dedicated to telecom; I and another for computing, system architecture and data center projects. In additional we worked together with other two engineering teams, deployment and operation. It was a great time; we did more the 2.000 different and very successful projects. All projects were customized for each client needs, because of that there was not daily routine at work. When we see a project done and working in production, we have a good feeling of mission accomplished. Our team got awards every year regarding the accomplishments. I got motivated to start my own business with this experience: dealing with clients, designing custom solutions, doing project management, costs and pricing.
I am working in my own startup, A2F, founded together with two friends. We work designing custom projects in two main areas:
Brazil is in a good shape! All kind of business with technology could be very successful for the next years. In a short time, our company got interesting projects, partnerships and contracts. We will work very hard the following years because there is a lack of engineers and IT specialists. There are general needs at companies for: automation, system integration, improvement of quality, reduces human errors, get green, to be more productive (faster), get knowledge from systems and do more spending less money. This is exactly the kind of issues that can be addressed by technology applied to business.
Thinking out of the box, getting more creative, seeing how technology can help people lives, improving business and be eco. With the lack of STEM professionals at the market place and the number of growing projects, we really need to build up our own tools, to automate our work in IT and engineering, and do more with the same amount of people, keeping and improving the quality. Give the dirt job to the computers. Furthermore, I usually try to motivate high school students with good grades in math and physics to follow Engineering or Computer Science career; those are challenge fields to work, with good perspectives for the future.
My twitter is @danielscarvalho