Blog

25/01/24 - Statistical Finite Element Modelling for Misspecified SST Simulation and Inversion

My first PhD talk! If you wanted to see a part of my PhD, this would be a good place to start. Thanks to the Complex Systems Seminar group.

08/01/22 - Data Science at 4CDA

Had the chance these past months to work with the people at 4CDA as an intern on a variety of problems related to Data Science, Visualisation and Consultation. While I have to keep the details light, I can say with confidence that it was a real deer in the headlights situation for me, getting to experience solving real world problems using the skills I had accrued in my education so far. 

Rather than just talk about what I did, or what I felt- I thought it would be good to hone in on that feeling I experienced and note down what I felt most surprised with, alongside how I dealt with it.

1. Time, time, time!

• A full time job really does mean 'full time', huh? You spend your time at school, about six hours a day of probably good concentration, at university, maybe a little more, and then you are expected to move into full time work. Nothing wrong with that, you need to get the hours in to get paid, right?  What is shocking really though, is just how long eight hours a day, five times a week can be! I mean seriously, in a problem solving job like data science consulting, I could spend the time doing all kinds of things, but I cannot imagine doing mental labour- filling out spreadsheets, solving non-issues for that long. I was really surpised at how much time I had. My entire life has been ruled by deadlines so to speak, so to finally have time to do things just really caught me off guard. What was important though, was to structure my own time effectively. In order to properly switch on at 9am, and then off again at 5pm, while effectively communicating, hitting deadlines, and maintaining mental wellbeing, you have to micro-timetable your eight hours, and macro-timetable your week. I found it best to do this by answering questions like, "When are you going to want to not do work?", "When are you going to be in communication/email/talking mode and when are you not?", "What specific time are you the best at knocking down tasks?"

2. The Cohesive Product

• I think one of the things University and School can fail to accurately emulate is the idea of the cohesive task. Some assignments are good at capturing this kind of thought process, but none come as close as business. What exactly however do I mean when I say, 'The Cohesive Product'? Well, I am referring to the idea that you are not going to hand in the working, finished thing and call it a day.  You are going to bundle it with documentation, and with labels, and comments, and citations, and all of the metadata that makes a product robust. This was new to me. I was given a task, but when I held it up and yelled 'done!', I was met with a barrage of "have you done this?", that got me to recognise how serving up a minimal product in the shortest time possible might not be as good as serving up something more cohesive alongside commentary. I remember bundling my data analysis toolset with a markdown document, and a front end GUI that made context descriptions between the various components more clear.

3. Options

• Quite similar to the "The Cohesive Product" is giving options. For one of the major tasks I ended up completing I gave four completely isolated methods for answering the problem and packaged them together. While in a University assignment I probably would have given what I thought was the best, in a business environment, when you hand over a product to someone else, you can't tell what context it will be used in. Therefore, giving as many options as possible puts more decisions, power and responsibility onto the client, which in the business world (at least from my understanding so far...) seems to be a pretty good move, as it puts less pressure on yourself to solve things later down the track.

So yeah, many, many lessons learnt. A lot of character building for me personally. What a fantastic experience though, to be treated like a real consultee, and given so much freedom to explore solutions to problems. I did manage to catch covid though...

12/03/22 - Ready, Trader, Go!

For a while now, investing has been an interest of mine. Anybody who has taken a high enough level high-school math course or an intro finance course will know the keyword- Compound Interest. It stood to reason in my 18 year old mind that if I invested all the money I had ASAP, I would see big gains down the line. 

I have seen money come and money go since. I definitely don't gamble enough to play the options game, and I'm not so held back I only invest in ETF's, or Exchange Traded Funds- the investing equivalent of a comfortable two decade old couch you can't climb out of. But on top of that, I have since learnt a lot about computer algorithms, and time series analysis- two fields I thought would put me in a place to start learning about algorithmic trading.

One of the largest firms in this space is Optiver, who as far as I can tell, utilise a fast infrastructure to execute a neutral, pairs-trading strategy. They are obviously always on the lookout for new talent, students out of university who are still neuro-plastitic enough to apply and execute advanced and innovating algorithms. To that end, they have a competition called, Ready, Trader, Go! which I've been fortunate enough to be apart of.

The competition is still going too, the first round (which unfortunately I was too busy to be a part of) finished on the 10th. It has been a real learning experience.

Part of the frustrations of mathematicians all over I'm sure is computational complexity vs. mathematical complexity. A good example of this is matrix multiplication, which high school students can do fairly easily if you tell them the rules, but computers must implement by more complex means so as to not be slowed to a crawl. 

So far in this competition, I have had many ideas I wanted to implement. But, I have so far been unable to execute on any of them, as the underlying programming necessary to recreate them has eclipsed my programming ability. My frustration grows, but perhaps if I finish my work next week, I'll be able to figure it out, come back here, and discuss my triumphs. 

Regardless, the competition is definitely a good experience, you get to work with a real business codebase, which to someone who has only ever programmed for myself, has been terrifying- but as with most terrifying experiences, a fantastic opportunity to learn.

04/03/22 - Always Carry a Speakable Piece of Maths.

Early January, I saw a post in the Facebook group >implying we can discuss mathematics about a mathematics blog aimed broadly at students in university. Being part of the target demographic, I obviously began investigating and found a whole array of useful and fun stuff to pick apart. Examining further I found a particularly useful post titled, Get a Maths Job which piqued my interest even further. I even went so far as to contact the author, Adrian Dudek, and found that he was the editor of the entire blog! Fast forward a bit, my blog post I posted for submission as part of the AMSI VRS was accepted to his website, and you can find the link here. 

It's titled, Always Carry a Speakable Piece of Maths and focuses on why it can be so hard to be a mathematician in a world of people who despise their middle-school math experiences.

Just goes to show you how a simple email can get gears moving, all you have to do is press send.