The Outdoors

A high-altitude traverse across the perovskite composition landscape. The climber maps 43,000+ devices across a 410-feature face, learning which chemical handholds (composition, fabrication) lead to stable footing at the summit (long-term device stability). Temporal splits keep the ascent honest.

A short, elegant problem on a small wall: 20 aryl halides, hand-built by the climber, interrogated for what their 2D fingerprint reveals about their reactivity with palladium. The route is intimate and exploratory — less about the summit, more about reading the rock.

A high-exposure ridgeline comparing two views of the same mountain: the slow, rigorous DFT path (ORCA) and the blazingly fast ML interatomic potential path (Meta's UMA/eSEN). The climber benchmarks Gibbs free energies and transition state barriers, asking whether the new-school route gets you to the same summit as the old-school one.

A route that ventures off the well-mapped face of ordered protein domains and into the loose, unpredictable terrain of intrinsically disordered regions. The climber asks whether the handholds that work on solid rock (structure-derived features, pLDDT confidence) become useless on the crumbling IDR face — and whether mutation effect scores follow different distributions in disorder.

A focused drug discovery route targeting the SARS-CoV-2 main protease. The climber navigates 2,000 compounds through a fingerprint-based ML pipeline, asking which molecular shapes look like good inhibitors. A Tanimoto similarity anchor to a known inhibitor (Nirmatrelvir) provides a chemical compass for the ascent.

A two-pitch route connecting OpenMM molecular dynamics to machine learning. Pitch 1: run MD simulations on xenon cubes, dipeptides, and small peptides, extracting energy data across simulation frames. Pitch 2: train an ML model to predict peptide energy from structure, using dipeptide simulation data for training and small peptides for testing. The route asks: can we learn a surrogate model for molecular energy?

A clean, well-scoped drug discovery problem: predict BACE-1 inhibitor activity from Morgan fingerprints, comparing a logistic regression baseline to a Random Forest. The scaffold-based split stretch goal adds a chemically meaningful generalization test — can the model recognize a new scaffold it has never seen?

A high-stakes route connecting the invisible chemistry of exhaled breath to the clinical reality of cystic fibrosis. The climber navigates hundreds of VOC features using Random Forest and TabNet — a deep learning architecture with built-in attention — to find the handful of biomarkers that distinguish infected lungs from healthy ones.

A technical traverse across the conformational landscape of cyclic peptides — molecules that change shape depending on their environment (aqueous vs. membrane). By computing 3D descriptors at two dielectric constants and correlating the difference (Δ) with experimental permeability, the climber attempts to quantify the 'chameleonic potential' that 2D fingerprints completely miss. This route connects directly to Jorge's thesis DEL platform.

A route through the emerging landscape of molecular glue degraders (MGDs) — small molecules that hijack the cell's protein disposal machinery. The climber navigates ~100 compounds from their own research, comparing Ridge regression on Morgan fingerprints to Chemprop (a message-passing neural network), predicting degradation efficiency (pDC50, Dmax) against both the target protein (ZBTB11) and an off-target (IKZF1). The route asks: can ML help design selective degraders?

A computational traverse through high-resolution mass spectrometry data from the polluted Tijuana River, hunting for PFAS — the 'forever chemicals' — in surface foam. The climber extracts spectral features, screens against 4,777 suspect compounds (NORMAN database), and uses Kendrick mass defect analysis to identify homologous CF₂ series. The route connects to real environmental health research in the Prather lab.

An ambitious infrastructure route: building a RAG-powered chatbot for quantum entanglement Q&A using Deepseek R1 distill Llama 70B (4-bit quantized) on an RTX A6000. The climber fetches scientific literature via the Springer Nature API, converts XML to JSON, and implements retrieval-augmented generation with a Dockerized deployment.