Sohhom Bandyopadhyay

quantitative research  /  cognitive science  /  applied machine learning

I studied computer science to understand how computers work, moved to cognitive science to understand how minds work, and now study financial markets to understand how money works. The through-line is measurement: I am most useful where the hard part is deciding what to measure, and then working out whether the number that comes back is real.

My PhD asked what active noise cancellation does to the music you actually want to hear — a question the field had largely skipped in favour of measuring the noise it removes. Answering it meant building a modulation transfer function for a headphone, pointing a computational auditory-nerve model at consumer hardware as a measuring instrument, and putting the result to eighty listeners. The headline effect turned out to be small, and writing that up honestly was harder than producing it.

These days I build trading systems: data pipelines, feature and label engineering, backtests that survive transaction costs, and decision rules that know when not to trade. Same discipline, different domain — a backtest is just an experiment with an unusually motivated way of fooling you.

• quantitative finance
• market microstructure
• psychoacoustics
• signal processing
• applied ML & LLMs
• experiment design