AI-driven drug discovery
Screening billions of compounds against AlphaFold-derived receptor ensembles, with active-learning triage and ADMET/QSAR profiling to cut wet-lab burden.
- AlphaFold3
- Boltz2
- AutoDock-GPU
- RDKit
Computational biophysics · Stanford Medicine
PRITAM
Panda
I build the computational machinery behind new medicines — from quantum-scale physics to billion-compound screens to the pipelines that keep it all reproducible. Where atoms move, life evolves.
NOISE
2.5B COMPOUNDS
Scroll
Open to collaboration
●AI-driven drug discovery●Molecular dynamics●Free-energy perturbation●AlphaFold3 · Boltz2●Multi-omics●Nextflow · Snakemake●GPU / HPC●Agentic scientific tooling●Protein design●Variant calling
The short version
Quantum physics taught me rigour. Biology gave me a reason to use it.
I am a postdoctoral scholar in Anesthesiology, Perioperative & Pain Medicine at Stanford University School of Medicine, where I design battlefield-viable anesthetics using AI-driven protein design and high-throughput virtual screening.
Before Stanford I built multi-omics infrastructure at the German Cancer Research Center, modelled antibody clones at Karolinska, and completed a PhD in physics at Uppsala bridging DFT-level material simulation with atomistic biology.
49
h-index
6,300+
citations
89
Q1 publications
1
US patent
What I work on
Four things, done properly.
Atomistic simulation
GPU-accelerated molecular dynamics and alchemical free-energy workflows that turn docking scores into binding affinities you can act on.
- GROMACS
- OpenMM
- FEP
- Metadynamics
Reproducible infrastructure
Production Nextflow and Snakemake pipelines across HPC and cloud — multi-omics, variant calling, and virtual screening that run the same way twice.
- Nextflow
- SLURM
- Docker
- AWS
Agentic scientific tooling
ReAct agents that drive real bioinformatics tools, with evals and failure-case curation instead of demo-grade prompt chains.
- Claude Code
- Tool-calling
- REST
- Next.js
Latest
Recent work & notes.
Loading…
Got a molecule, a dataset, or a pipeline that won't behave?
I take on collaborations across drug discovery, structural biology, and reproducible research infrastructure.