Bruno Fernandez-Ruiz
mailto:brunofr@olympum.com
http://www.olympum.com/
Linkedin Profile
Software architect of internet-scale application, interested in emerging and disruptive technologies. Depending on the hour, I play the visionary technologist or the nuts-and-bolts software engineer. I enjoy working with forward-looking, pragmatic and highly motivated teams across engineering, product and business. I prefer to work in fast iterative partitions in any given problem space and I continuously strive for simplicity. I am a strong believer in technology as a business enabler, and I take most joy out of working on creative proposals and solutions that address known and unknown problems in unique and original ways.
Working on the Yahoo! Platform on search, content, analytics, video, mobile, social, geo, local and membership.
Reporting to the SVP for International Engineering, I was responsible for setting up the architecture practice from the ground up outside of the US and linking up with the US groups. I built a network of 25 architects worldwide, defined engineering standards and design patterns, and achieved reviews on 100% of major projects. I also participated in the ASC/ESG, CAA and YArch platform reviews, promoted the use of platforms in international, initiated projects to ensure availability of development environments in all properties, and produced a prototype for a cross-property linked-data content store.
Transparent IM proxy for mobile devices. Acquired by leading messaging solution vendor.
I was responsible for operations architecture across Fidelity International in 3 locations and working with 800 engineers. I led the definition and maintenance of a 5-year strategic technology roadmap classifying technology as emerging, tactical, strategic, containment, retirement, or obsolete. I worked heavily on OS virtualization, ensuring applications would be fit for a virtualized environment, and putting in place the necessary infrastructure. This entailed switching to diskless blade servers, introducing automated OS and application provisioning, adopting Network Attached Storage in production for blades, and establishing application-level production monitoring. I also put in force the architecture governance, led the Architecture Review Board in FIL, and I was part of Fidelity's company-wide Architecture Review Board.
Yamigo was a startup offering mobile instant messaging services. The solution was based on the Wireless Village initiative, later integrated into the Open Mobile Alliance Instant Messaging and Presence Services. Yamigo had over 150,000 registered users and 15,000 active users per month.
State Farm is the world’s largest P&C Auto insurance company. In 2004, I was part of the business development team that sold Accenture's CCS for J2EE to State Farm’s CTO. Consequently, I worked with a team of 25 engineers making changes to the development architecture in order to increase developer productivity. These changes involved refactoring the application to reduce coupling, increasing test coverage, improving the Platform-Specific Model (PSM) to ensure changes could only be done on the UML model, and introducing new a classloader for hot loading the PSM generated classes during development.
Accenture partnered with Topdanmark, a leading Nordic P&C mutual insurance, to develop the first charter client implementation of the Accenture’s Claim Components Solution (CCS) for J2EE. I was part of the business development team and after the sale I became responsible for jump-starting the project, day-to-day client engagement, and for designing the integration architecture against the existing IBM IMS and CICS systems.
I was responsible for re-architecting Accenture's Claim Components Solution (CCS) from Microsoft's .NET to J2EE, playing a combined role of product manager and architect.
I run the final performance benchmark of the .NET and J2EE applications at IBM’s Benchmark Centre for Microsoft technologies, and the results showed comparable scalability and response times in both platforms.
I led a team of 7 to ensure the production readiness of the Meridea solution in Sampo’s environment, including F5 BIG-IP, IBM Tivoli PolicyDirector, IBM Http Server (IHS), IBM WebSphere Application Server (AIX), IBM DB2 (z/OS), IBM DB2Connect, IBM CICS, and IBM MQSeries (z/OS). This project was critical for my development as an architect, giving me a deep understanding of the full end-to-end stack, including inner workings of JVM, WebSphere, and the Linux and AIX Kernels. Sampo was the first bank worldwide to run a seamless banking product experience across mobile and web channels on a single technical platform. This has allowed Sampo to reduce the overall IT cost per account by 300%.
I joined this project during the feasibility study phase of what was to become Meridea, a mobile banking software joint venture between Nokia and Accenture. I was initially responsible for the conception and prototyping, for which I developed an end-to-end mobile and web banking application for IPO subscription, leveraging Nokia’s mPlatform Operator Gateway and BEA Weblogic, and supporting the yet-to-be released Nokia 9210 (Symbian EPOC OS). The application was a complete success when shown in CeBIT’01, and a key leverage during the funding phase, which closed in raising €45 million with 3i.
I owned the overall concept and design, and worked with a team of 130 engineers (Philippines and Finland) building the underlying mobile application framework built on top of the J2EE platform:
After the first two releases of the software, I became responsible for overseeing all architecture decisions across all release streams, working together with the two development locations (Finland and Philippines).
I researched personalization techniques and its applicability to eCommerce. I reviewed commercial software implementations from BroadVision, ATG Dynamo, Blue Martini, Vignette and Calico, as well as developing new models based on collaborative filtering and utility maximization through a multinomial logit.
I remained affiliated with MIT's Intelligent Transportation Systems lab, providing technical advisory on research programs and proposals, as well as reviewing papers.
I integrated MITSIM and DynaMIT traffic simulators. MITSIM is the world’s leading parallel microscopic traffic simulator written in C++ and using the Parallel Virtual Machine (PVM). DynaMIT is a distributed real-time traffic prediction application. I enhanced DynaMIT to receive real-time traffic readings from MITSIM, and MITSIM to receive real-time travel information from DynaMIT, by introducing a PVM - CORBA bridging component leveraging on CORBA’s asynchronous messaging service.
I ported MIT's microscopic traffic simulator (MITSIM) to Linux, switching to GCC, libstdc++ and Lesstif.
I created a new front-end for policy administration as a rich-client in a Java applet, encapsulating the AWT into a component library that was later eventually replaced by SWT.
I developed programs in C++ using the finite elements method to calculate extreme conditions and failure on very large-span structures.
Studied operations research, applied to transportation systems. Specialized in discrete choice and random utility modeling. Nonresident doctoral thesis research.
Studied computational methods for systems engineering.
Specialized in computational physics.
Mono is a project to develop a C# compiler (ECMA 334) and Common Language Runtime (ECMA 335). I have contributed patches to the Garbage Collector, initialization thread, System.Xml and interop services. I also developed GtkGL#, a binding for the OpenGL libraries on a Gtk# canvas.
JBossMQ 2.x series was a clean-room implementation of the Java Messaging Service, including guaranteed delivery, persistent messages, and durable subscriptions.
Apache Xalan-J is an implementation of an XSLT Processor, implementing the JAXP API. I contributed to the XSLTC compiler and bytecode generation.
Caffeine is a free high-performing interoperability solution between the Java platform and the .NET framework. Unlike other solutions, Caffeine does not replace the Java Virtual Machine, but embeds a JVM runtime within the .NET Common Language Runtime process, avoiding IPC costs, and ensuring that all vendor-specific VM optimizations are preserved.
Piano, cooking, swimming, and catamarans.