About Me

me-Tux-whiteI am a frequent traveller, specializing in large scale databases. My experience includes: performance tuning some of the largest SQL Server installations in the world, doing benchmark runs and POC.

Full details of my skills can be found at LinkedIn. My biography and picture, which you are free to use if I speak at your conference, is available below.

 

Thomas Kejser Biography

Thomas started his career as a teenage coder, but moved into the data management space in the 90’es and never looked back. He has been modeling and tuning databases since and is passionate about perfecting his high scalability design skills.

He holds a masters degree in Computer Science from DAIMI in Denmark and has industry experience across a diverse range of sectors including  telcos, software houses, healthcare, manufacturers, retailers and investment banks.

Building multi-terabyte sized data warehouses, setting up OLTP systems for extreme scale and tuning world record data movement speed are some of Thomas’ recent achievements. If you need a bottleneck in your system removed, from code to hardware, then Thomas can help you.

Sharing of knowledge is important to Thomas. When he is not coaching developers and DBAs, he publishes whitepapers for Microsoft and his musing on http://blog.kejser.org.

Thomas lives in London and spends his spare time discussing philosophy with friends over a good glass of whiskey.

 

Curriculum Vitae

Below, you will find a history of my work life and education.

Blog: blog.kejser.org
Email: thomas@kejser.org

Publications / Presentations

I write whitepapers and publish them. I also review the work of others and contribute to their publications. Below is a selection of my work.

I am also a frequent presenter at conferences where I generally score among the top 10 best presenters. You can see an example of my presentation style here:

My work also includes teaching. I am one of the main contributors to the original Analysis Services Maestro Certification Program and I create teaching material on SQL Server for Microsoft employees. You can see an example of my work here: Designing I/O systems for SQL Server.

I also maintain my own courses, which can be found here: Courses.

 

Industry Experience

My jobs have taken me around many industries. Below are some highlights.

Industry Experience Highlight
Retail Performance tuned largest retail cube on the Microsoft platform achieving load speeds of 2M sales tx/sec. Tuned cube at largest Danish Supermarket and helped with 2000 to 2005 migration (DSG)Implemented cube for largest Nordic DIY retailer (DT Group). This project was mentioned in the Danish newspaper Børsen and has since become a reference case for Microsoft and HP. Design and implementation of purchasing and sales data warehouse for Valora Trade using meta driven data modeling.
Telecommunications Data modeling and troubleshooting on largest CDR record on MS platform at the time (75TB on single Superdome). Hardware tuning on Danish Telco Telmore. Helped build high speed, low cost I/O system. Achieving 6GB/sec table scan speeds.
Retail Banking Enterprise Data Modeling assistance at Danske Bank, including handling a 150M row customer dimension with historical data.Performance tuned retail banking model at Kuveyt Turk to achieve 13K credit card tx/sec on a single machine.
Investment Banking Risk analytics cubes loading data at 3M rows/sec and serving 1000 concurrent queries across 2 geographical locations at UBS. Architected and tuned risk analytics cube serving 300 concurrent queries at <1sec latency.
Healthcare Implemented data model, security system and ETL logic for data warehousing holding most patient records for the small country of Denmark. Migration of SQL Server 6.5 and VB5 codebase to modern platform at Skejby Sygehus using automated code analysis and resulting in a large performance improvement.
Software Frameworks Build metadata framework for auto generating  ETL and data models. Allowing extremely efficient construction of data warehouses. The framework is now a standard part of the Platon development process. Build simple, versioned object store as part of University project (The GAIA Framework)
Government Implemented unemployment statistics warehouse for the Danish Council of Counties for web based consumption.
Benchmarking Performance tuned SQL Server to achieve ETL World Record on a Unisys 64 core machine.Tuned SQL Server transaction log throughput to 950MB/sec on high speed NAND devices using L2 cache coherence techniques
Online Gaming Performance tuning on large superdome system for online gaming system eliminating bottlenecks and implementing high speed data movement and archival process.
Manufacturing Performance troubleshooting and architecture assistance on wind farm monitoring data warehouse with 100+ TB data. Advised on data model changes to achieve 3x compression over existing system.Geo distributed replication architect for Dandy chewing gum.Implemented financial consolidation warehouse for Bombardier transportation allowing translation between ERP systems.

Skills

It is generally hard to come up with a good, global taxonomy of skills. But if I were to try, I would highlight these:

  • Coaching and teaching
  • A zero BS attitude to life
  • Data Modeling of both OLTP systems and Data Warehouses
  • High scale performance tuning
  • Scientific, analytical thinking
  • Deep and wide understanding of software and hardware
  • Concurrent programming and data structures

On programming:I have worked with several programming languages throughout my career and find it easy to pick up new ones. Below are some examples of the languages I have written working and useful code in:

  • C# (my current favourite)
  • Ansi C and C++ (including using Lex and Yacc)
  • Perl
  • Visual Basic
  • Java
  • Motorola 68000 assembler
  • ML
  • Ruby
  • JavaScript
  • SQL (many variants)

Further information on my skills, using the LinkedIn taxonomy, can be found on on my profile.