Member of Technical Staff at Ditto
I do ML/AI engineering for startups.
My offline cascade took classification accuracy from 78.6% to 98% at $2.03 a game on H100s.
I placed in the 74th percentile at the Putnam, and I'm on track to graduate in December 2027.
Muon works on muon-catalyzed fusion, and our patent covers novel mechanisms for potential sustained fusion.
My role-aware LLM pipeline over 13+ annual SOPs hit 85% Recall@10 for the Massachusetts state government.
Fixing label circularity in my 25-year SPY market regime classifier gave an honest 0.876 out-of-sample AUC.
A three-stage LLM pipeline of Claude, Perplexity, and Gemini filters spurious pairs, for a 0.96 Sharpe out of sample.
Member of Technical Staff at Ditto
I do ML/AI engineering for startups.
My offline cascade took classification accuracy from 78.6% to 98% at $2.03 a game on H100s.
I placed in the 74th percentile at the Putnam, and I'm on track to graduate in December 2027.
Muon works on muon-catalyzed fusion, and our patent covers novel mechanisms for potential sustained fusion.
My role-aware LLM pipeline over 13+ annual SOPs hit 85% Recall@10 for the Massachusetts state government.
Fixing label circularity in my 25-year SPY market regime classifier gave an honest 0.876 out-of-sample AUC.
A three-stage LLM pipeline of Claude, Perplexity, and Gemini filters spurious pairs, for a 0.96 Sharpe out of sample.
By using this service, you agree to the Terms of Service and Privacy Policy.