Writing
Long-form notes, filed under four topics. Drafts are marked; they are published when they stop changing.
- FINExpected Shortfall vs. VaR: what the number actually promisesDraftVaR tells you a threshold; ES tells you what the bad days cost. Why Basel moved, why desks still quote VaR, and what I learned implementing both on the same pipeline.
- DSSpark 4 on a modern JVM: Java 21, Java 25, and the APIs worth caring aboutDraftWhat changed in Apache Spark 4, how it behaves on Java 21 and the Java 25 LTS, and the API additions that actually alter how a risk engine gets written.
- SWEReading CUDA release notes like a systems engineerDraftWhat actually changes between CUDA toolkits, which of it matters for latency-bound code, and a running log of the releases I've tracked.
- FINmarket-risk-quant: three VaR methods on one bookDraftA Value-at-Risk platform that runs Parametric, Monte Carlo and Historical Simulation over the same positions in the same job, reports Expected Shortfall alongside, and treats the disagreement between methods as the output.
- AILocal AI systems: harnesses, Argon, and why the benchmarks lieDraftRunning models locally is the easy part. The harness around them is the system. Notes on building one, where Gemini 4 Argon fits, and a critique of the benchmarks everyone quotes.
Nothing filed under this topic yet.