Geoffrey Jost
geoffrey at jostwebwerks.com
/
626-380-0806
530 S. Lake Ave #274,
Pasadena, CA 91101
US
I've been building and maintaining complex web applications for well over a decade.
What I Do
Custom Web Development
Everything from simple blogs to a large multilingual social networking site.
Linux System Administration
I've been running and administering various flavors of Linux (Debian, Ubuntu, Red Hat) for the last decade, and I've also done some work with FreeBSD.
Web Design
I'm not a designer myself but I'm pretty handy with graphic design tools. If you have an idea I may be able to execute it. If not I can work with designers who can.
Free/Open-Source
I believe that using free and open-source software and standard data formats is in the best interest of my clients.
Current Work
Densho Digital Repository (2011-present)
- Densho collects and preserves historical materials in digital form for itself and a number of smaller organizations.
- Partner organizations are geographically distributed, often with poor connectivity. The materials collected are archival-quality and voluminous (e.g. uncompressed digital video). The goal is for these materials to be preserved for the long term (e.g. longer than a hundred years) making a traditional centralized CMS impractical.
- The editor stores metadata in text files in Git repositories. Binary files are synced between collections by git-annex via Internet or sneakernet.
- The public site consists of a Django front end with data in Elasticsearch.
- The DDR toolset includes numerous command-line tools for importing, exporting, and otherwise managing archival data as it moves towards publication.
Densho Encyclopedia (2011-present)
- The editing back end is a MediaWiki instance, access to which is restricted to staff.
- The public site consists of a Django front end with data in Elasticsearch.
Past Work
Japanese American National Museum (2003-2011)
- Implementing customized mini-sites for major exhibitions.
- Providing tools that enable non-technical staff and users to manage web site content.
- Administering a cluster of Linux database and web servers, including development and staging servers.
- Wielding Python and any other useful tools to process data as needed.
- Developed JANM online collections, a Django front end to the Museum's back-end Argus collections database.
DiscoverNikkei.org (2003-2011)
- Community site with interface and content in English, Japanese, Spanish, and Portuguese.
- Site includes the Nikkei Album, a multimedia storytelling application.
- DiscoverNikkei.org awarded Best of the Web at Museums and the Web 2007.
Los Angeles Metropolitan Transportation Agency (2007-2008)
- Did back-end Django programming for numerous Metro.net site properties, including Service Advisories, customer feedback processing (now offline), the "Where Does Metro Take You?" itineraries contest, and for the promotional banner graphics used throughout the site.
- Architected dev/stage/production server environment with version control using Git.
- Led transition from Apache2/mod_python to Nginx / Gunicorn with Memcached.
- Prototyped REST back-end for a new Trip Planner interface using web.py.
- Assisted with Red Hat Enterprise Linux server administration.
Vivendi Universal (1999-2003)
- Designed and developed custom Java/JSP infrastructure, including user administration and content management and workflow tools.
- Implemented XML-RPC back-end for an early American Idol tie-in game.
- Assisted implementation of payment processing system using CyberSource.
- Trained colleagues in HTML, web graphics production, and usability issues.
- Administered the site's Apache2 and Tomcat setup.
- Managed project source code with CVS and (later) Subversion.
Knowledge Adventure (1999-2003)
- Development support for various products, including JumpStart, Math Blaster, and ClassWorks (when the latter was at KA).
Tools
My current tools of choice are Python and Django but in the past I've used PHP, Ruby, and (a long time ago) Java and Perl. I've worked with many of the standard web-development tools: Nginx, Apache2, Memcached, Redis, MariaDB (MySQL), and the venerable PostgreSQL. For quick and portable database fun it's hard to beat SQLite! More recently I've used Elasticsearch as a document store and search engine. I like to use automation tools like Ansible to keep things solid, and of course I manage my source code with Git.