Updated REPLACE_DATE
- Building Telos, most days. The current problem is that semantic similarity gets you a plausible candidate and placement history tells you whether a recruiter would actually submit them — and those two signals disagree more often than I expected. Recent work: pushing candidates back into JobDiva, one-click outreach through Microsoft Graph, and comparing a person against who’s already been submitted to the req.
- Learning CS at Harvey Mudd through the 5C consortium. Also working through NeetCode properly rather than pretending I already know it.
- Research Still chewing on the Kalshi lead-lag result — a two-day edge on unemployment markets that vanishes entirely on CPI. I don’t yet have a story I believe for why the two differ, and I’d rather say that than invent a mechanism.
- Reading REPLACE_READING
- Training Half marathon. Breaking, still competing. Guitar on Monday nights at Pomona.
- Open to Talking to anyone working on agent reliability, labor-market data, or calibration. rushiljaiswal1@gmail.com · rjaiswal29@cmc.edu
~/R