Luke Hornof
About
Luke Hornof is from San Francisco Bay Area. Luke works in the following industries: "Computer Software", "Semiconductors", "Internet", "Higher Education", and "Information Technology & Services". Luke is currently Co-founder & CEO at Luminide, located in San Francisco Bay Area. In Luke's previous role as a Director, AI Product Research and Marketing at Intel Corporation, Luke worked in Santa Clara, California until Jan 2020. Prior to joining Intel Corporation, Luke was a Director of Software, AI Products Group at Intel Corporation and held the position of Director of Software, AI Products Group at Santa Clara, California. Prior to that, Luke was a Director of Software, Nervana at Nervana Systems, based in Mountain View, CA from Aug 2014 to Aug 2016. Luke started working as Enterprise Integration Engineer at Pandora in Oakland, CA in May 2014. From Feb 2012 to Jul 2013, Luke was Senior Software Engineer at Root Music, based in San Francisco. Prior to that, Luke was a Senior Software Engineer at AdBrite, based in San Francisco from Nov 2009 to Nov 2011. Luke started working as Senior Software Engineer at Sun Microsystems in Silicon Valley in Apr 2004.
Come check out Luke Hornof's email address on finalscout.com, a free professional database with 500 million business professionals and 200 million companies.
Luke Hornof's current jobs
Luminide accelerates AI innovation with a focus on model development.
Luke Hornof's past jobs
Hosted the NeurIPS 2019 Intel AI Research Luncheon, produced demos for the 2019 Intel AI Summit, technical advisor for the Tokyo 2020 Olympics.
Built/managed a 150 engineer AI SW team (US employees: 30, Poland: 60, India: 50, Israel: 10) ・Strategy, vision, culture, hiring, global scaling, roadmap, execution ・SW/HW Architecture, Algorithms, Compilers, Simulators, Distributed, Performance, Driver/Firmware
(member of Playground Global in 2015, acquired by Intel in August 2016)
Designed and implemented backend server code for a system that supported 500,000 artists and 30 million MAU, including REST and graph APIs, artist search using Solr, and Amazon messages queues. Integrated with 3rd party APIs such as Facebook, Twitter, and CrowdFlower. Key technologies: Java, Spring, Hibernate, MySQL, XML, JSON, Eclipse, git, maven.
Designed and implemented software for the largest independent ad exchange, which serves over 150 million impressions a day on over 60,000 sites. Worked on core server components such as inline, XML, and real-time video ads. Used predictive models generated from Hadoop map/reduce. Built a number of publisher, testing, and ad review tools.
Designed, implemented, tested, and debugged software for an advanced video server, capable of storing 300,000 hours of content and delivering 160,000 2 Mbps streams. Also built a large portion of the user interface and most of our end-to-end testing infrastructure. Technologies included Linux and Solaris, TCP/UDP sockets, and shared memory. The C++ code base used templates, design patterns (e.g. smart pointers, reactor/observer), multiple threads, and a 64-bit address space. Debugging memory leaks, dangling pointers, and threading errors was done with tools like gdb, dtrace, valgrind, and tcpdump. Agile software strategies included daily stand-up meetings, test-driven development, pair programming, and continuous integration.
Joined Andy Bechtolsheim and David Cheriton's latest startup. Worked on designing, implementing, and testing software for the next generation digital TV. Kealia was acquired by Sun Microsystems in April 2004.
Implemented and tested an interpreter in C for the AMD-64 instruction set. Delivered product on time, despite having an extremely ambitious deadline. In addition to excellent software engineering skills, this project required intimate hardware knowledge and attention to very low-level details. Also worked on an abstract interpreter in Java to validate x86 translations.
Created "self-specializing mobile C code" that travels over a network and automatically optimizes itself to its destination. This technology required a breakthrough in type-theory which allowed the verification phase to prove safety properties about code not yet generated as well as run-time code generation to quickly produce highly optimized code.
Classes and research.