A $100M book, a real mandate, and one merger nobody else sized

A $100M simulated book run to a real mandate rather than an unconstrained optimiser. The constraint was the interesting part: an ESG screen, a benchmark to beat, and a risk budget that made concentration expensive.

I did not start from the answer. The framework was built by testing equal weight, then mean-variance optimisation, before settling on an ESG-constrained barbell. That structure was chosen because conditions were moving from AI momentum toward geopolitical risk, and a barbell survives that rotation better than a single-factor tilt.

The final book returned 9.4% against a 1.8% benchmark at a Sharpe of 2.04 and a beta of 0.35. The low beta matters more than the return: most of the excess came from selection rather than from simply taking more market risk.

I spotted the Cyclerion and Korsana merger from public pre-market information and sized a small diversified position. It produced roughly $5.4M on a 205% opening-day move, the single largest contribution to the portfolio.

The sizing is the part I would defend in an interview. A 205% move is not something you can forecast with confidence, so the position was sized to be survivable if it went the other way, not sized to the upside case.

Testing three allocation methods before committing to one, and writing down why the third was chosen. It is slower, and it is the reason the result is explainable rather than lucky.