I am currently working remotely as a trading analyst for a Swiss commodity hedge fund, looking at shiptracking flows, weather & supply/demand data to extract trading signals for energy & ags. These lower frequency fundamental signals were to provide context to CTA-style strategies. 🛢️⛈️🌾
I am also working on various small incremental aspects of the codebase - systematic scraping, live market data processing, job scheduling and profiling.
I landed this role by reaching out on X/Twitter, showing my projects/portfolio with an earnest desire to learn, and found a willing mentor who agreed to let me work at his firm. To that, I am eternally grateful for the kindness he showed me, and I strove to repay it with hard work and grit.
As an analyst, I was responsible for the daily price assessment process for middle (jet fuel, diesel) and heavy distillates (marine fuel). I had to learn the fundamentals and technicals of these markets, the product specifications, the end-to-end process from extraction, transporting to a refinery, then selling the distillate and the specifications/mechanisms of the various contracts. Actually that's cap, I managed to learn none of that while there. I learn it by myself now.
Using Python to automate data parsing and price assessments, I implemented price assessment algorithms of daily index values for paper and physical markets with constraint programming/constrained optimization, resulting in a more objective assessment process with time savings.
As a data science intern, I created topic modelling/keyword extraction pipelines to extract investment themes from reports to assist portfolio managers in making investment decisions. This was done with classical and deep learning techniques.
I also created dashboards to track key performance and usage metrics of GIC's various databases to determine usage patterns.
I interned at proptech firm called Real Estate Analytics (8PROP). This firm was a property analytics provide that provided residential property valuation and analytics services via web-app/API that real estate developers, agents and investors could subscribe to.
The bread and butter of my internship was to tune the regression models that predicted property value which was the main product of the web app. This included tweaking the existing codebase, trying out different regression models, brainstorming and testing additional features against historical prices. I also used SQL for adhoc queries and extracting data.