Now
What I'm working on right now.
A short, dated note. Updated when the answer changes.
As of April 2026
01
Building
- → Standing up the in-house Snowflake / dbt / Streamlit analytics framework for Public Investments at Northwestern Mutual — moving the team from one-off notebooks to a shared, governed, end-to-end stack.
- → Hiring quants and analytics engineers who can write production code and talk fluently to portfolio managers in the same week. The bar is the intersection, not either skill in isolation.
- → Working on a small library of internal ML models that sit close to live decisions — relative-value, exposure attribution, and rebalancing analytics — and on the tooling that makes them reviewable rather than mysterious.
02
Thinking about
- → How investment teams can use frontier AI without losing the fundamentals — what gets delegated, what stays human, and what the next generation of buy-side workflows actually looks like.
- → The right shape for an analytics team inside a large investment manager. Centralized? Embedded? Federated? My current answer is "all three, deliberately" — and the deliberateness is the hard part.
- → Where the line is between a model that helps a portfolio manager and a model that replaces them. I think the line is much further from the manager than most vendors suggest, and the work is in proving it on real money.
03
Reading
- → Annual reports, conference call transcripts, and the kind of long-form sell-side research that still exists if you know where to look.
- → Papers and posts on agentic systems, evals, and the operational realities of running ML in regulated environments.
- → A long backlog of books on the history of asset management — the institutional memory most quants skip and then re-learn the hard way.
Inspired by Derek Sivers's /now page movement. The full reference document lives at /resume.