About

A team member is applying to the Two Sigma PhD Fellowship. We have a ≤500-word draft that must be ML-forward and aligned to Two Sigma (nonstationary, multimodal time-series; self-supervised reps; drift/adaptation; uncertainty). You will: Restructure the draft (problem → ML contributions → validation → Two Sigma relevance). Make the methods technically sound for ECG/EMG + sweat biomarkers (noisy, asynchronous data). Add concrete metrics (AUROC/F1, ECE/coverage, time-to-detect drift) and a short applicability paragraph to quant research. Keep to ≤500 words; clear, reviewer-friendly writing. Provide one revision. Deliverables: Final ≤500-word proposal (clean). Tracked-changes version with brief edit notes. 2 alternative titles + 5–7 keywords. Must-have: PhD/ABD in ML/CS/ECE or equivalent industry. (MSc or excetional BSc candidates are also encouraged to apply) research. Strong time-series ML background (rep learning, probabilistic/uncertainty, online/drift, multimodal fusion). Evidence of rigor (publications—NeurIPS/ICLR/ICML/AAAI, IEEE/ACM venues—or equivalent work history). Excellent technical writing. Nice-to-have: Wearables/biomedical signals experience. Budget & timeline: Fixed price $100. Deadline: 2 days from hire. How to apply (short): 3–5 bullet critique of our draft’s ML framing & Two Sigma relevance (≤120 words). 1 sample paragraph you wrote for a research proposal/methods. Name one calibration metric and one drift-detection method you’d reference.