Very old picture.
A long time ago, I looked like this, having the good fortune to work with Dan, both much younger, both doing PhDs in the department.

I took a commercial road, and Dan took the academic. Look where that got me!


First stop after the PhD, well, I was still writing up, but anyway -- I went to Zarlink Semiconductor, working for an old friend (I'd been in the incestuous technology pool of the inmos and derivative crowd so knew how flat the world was in the technology space of the south-west). Besides real work, if you might call it that in the lab in Caldicot, I wrote Edge-TV that still makes me laugh. I worked with some real fun guys.

Three and a half years later and looking at the company's fourth or fifth iteration of redundancies (this was early 2004), I was faced with a choice: Take my accrued annual holiday or lose it. I got called while off and met the early-state ClearSpeed. I say early in the software stack sense. This was pre-CSX600. Anyway, my work/time was interesting in the liberty of scope. I have a good collection of friends from the four years I spent there and probably the most bonkers bit of code was the Monte-Carlo Toolbox that had marketing hopping mad, probably because I'd picked a name without consultation. The other stuff was more industrious -- I'm not writing a CV here. Oh, to keep the incest going, I was of course, working for/with some old friends.

The last year and a half was something of a ride in a similar fashion to the demise of Zarlink. However, rather than be simply fascinated at observing the collapse of a company, remaining as a witness on the indside with little worry about meeting bills (oh, the days!) I had real needs to meet (baby, wife post-surgery, mortgage in that order). We parted company without any hard feelings on my part and I have plenty of ex-CSD pals and I think most of those that remain would still answer the phone if they knew I was calling.

From the fire, I jumped to the frying pan. A one year-ish stint at the other HPC firm in Bristol, Quadrics. A company with a product that certainly had a market, alas perhaps, wanting in other regards. Oh, and working with another old friend [for one month -- before he scarpered]. The product and work (besides the environement) itself was so niche (clique) that it was an amazing experience. To tell anyone that I wrote an Elan5/QSNetIII Inputter from scratch that worked at speed against the (bug fixed) DMA Proc in the last few months would take an age to explain. Some really cool people worked there before it properly went bust with the meeting to end all meetings.

Fortunately for me (I suppose), I was already serving my notice to go for a new path. Career two point oh. I'm now a Civil Servant. One of the things I am earmarked for is understanding a rather big computer. Bigger in core count than the mighty Titech TSUBAME or node-count of the TATA/CRL EKA machines I played with while at ClearSpeed. This one's not all in one room. This one's left to the reader: A twenty thousand (less change) ~2GHz node linux computer, ~10TB RAM, running about thirty million transactions per year. Installed (after much delay) through 2005/2006. Most everyone you know has used it (by-proxy), once a year since then.

Oh, of course, I'm working with someone I know. My boss is an old school friend from, errr, twenty years or so ago. The only gig I've done on a limb not knowing anyone before I started was De La Rue. That was an interesting place! I met a few wonderful people there.


James Irwin, lastnamefirstinitial --> gmail. Last modified on August 2009. © 2009 University of Bristol
.