Qiankang (Kant) Wang

GPA: 3.82 / 4.00

Coursework: Machine Learning (CS 189), Probability (STAT 134), Principles and Techniques of Data Science (DATA 100), Computational Molecular and Cell Biology (BIOENG C131), Data Mining (DATA 144), Natural Language Processing (EECS 183).

• Developed a framework for testing whether AlphaGenome and ChromBPNet learn the same regulatory features, comparing internal representations, base-level attributions (DeepLIFT/DeepSHAP), and motif instances (TF-MoDISco, FiNeMo) on K562 and GM12878 data, with in silico perturbation analysis as an orthogonal check.
• Found that the two models recover largely the same regulatory motifs, with only a handful of low-support motifs unique to either model.

• Implemented conjugate gradient (CG) and biconjugate gradient (BiCG) solvers in AmberTorchPB, a LibTorch-based framework for Poisson–Boltzmann reaction field energy calculations on CPUs and GPUs.
• Benchmarked accuracy, runtime, and memory use across numerical precisions, showing that the CG solver reproduces AMBER PBSA energies (R² = 1.00) while running over 2× faster on CPU and nearly 2× faster on GPU [1].
• Built a Slurm-based scheduling pipeline that executed over one million PBSA energy calculations.

• Implemented SimCLR-style self-supervised learning with a ResNet-18 encoder and an alignment-and-uniformity objective in PyTorch. Pretrained on 100,000 unlabeled Tiny-ImageNet-200 images for 45 epochs, achieving 26.2% top-1 accuracy on the 200-class validation set using frozen encoder features and a cosine k-NN classifier (k = 20).

• Developed a conversational LLM agent with scoped memory and context-aware decisions about when to respond. Built an asynchronous pipeline that records user feedback, requires corroborating evidence before applying behavioral updates, and supports audit trails and rollback; integrated messaging platforms through a shared gateway.