Ryan Kim

Designed an LLM autoresearch agent that drove 37% of the gain over baseline on MedARC's NanoPath v1 benchmark (3rd-author paper, NeurIPS 2026 ASCI), using dual-branch search with anti-overfitting resets. Architected the agent's research infrastructure (knowledge graphs, git-worktree isolation, preemption-robust SLURM jobs, run tracking), cutting config errors 80% and tokens 15×. Introduced SPECTRA, a label-free contrastive LoRA fine-tuning method for domain-shift robustness, raising mean PathoROB robustness 0.726→0.891 and downstream AUC (38 CPTAC tasks) across 7 pathology foundation models; first-author paper accepted at the NeurIPS 2026 ASCI Workshop.

DNABERT-based prognostic model for glioblastoma survival via Cox regression on non-coding regulatory DNA mutations from 185 patients; identified age-related survival signals and racial disparities encoded in mutational scores.

Benchmarked 12 pathology foundation models across TCGA and internal cohorts, finding a 2x increase in out-of-distribution tumor-purity error (MAE ~0.10 to ~0.20) and highlighting the resulting clinical deployment risk.

Led engineering for a high-concurrency math competition platform serving 150k+ users across 11 partner organizations, building the React/Node.js authentication, submission, and leaderboard workflows. Raised $20k in funding (Hudson River Trading, Innovate901 First Place).

Built a CPU-only, locally-deployed LLM pipeline (LLaMA 3.1-8B, Qwen3-4B, DSPy/MIPROv2) to classify 370 clinically annotated lung cancer CT reports; co-authored a referral-event analysis abstract accepted at the ACRO 2026 Summit. Applied Cox regression and survival analysis to 230 patients' records, linking percent thymic tissue to NSCLC outcomes.