Biomolecular design through latent reasoning with RFD4-Proteina
Preprint (2026)
Tomas Geffner
I am a senior research scientist at NVIDIA, working on generative modeling within the Fundamental Generative AI Research Group.
I work on probabilistic machine learning, focusing on generative models and sampling methods. I am interested in methodological developments as well as applications in different scientific domains.
Publications
2026
Large Language Continuous Diffusion Models
arXiv (2026)
Latent Generative Search Unlocks De Novo Design of Untapped Biomolecular Interactions at Scale
bioRxiv (2026)
DiLaDiff: Distilled Latent-Augmented Diffusion for Language Modeling
arXiv (2026)
Scaling Atomistic Protein Binder Design with Generative Pretraining and Test-Time Compute
ICLR 2026 (oral, top 1.1%)
La-Proteina: Atomistic Protein Generation via Partially Latent Flow Matching
ICLR 2026 (top 0.01% by average reviewer score)
2025
Latent Target Score Matching, with an Application to Simulation-Based Inference
Machine Learning and the Physical Sciences Workshop (NeurIPS 2025)
Consistent Synthetic Sequences Unlock Structural Diversity in Fully Atomistic De Novo Protein Design
AI4Science Workshop (NeurIPS 2025)
Learning Straight Flows by Learning Curved Interpolants
DELTA Workshop (ICLR 2025)
Efficient Molecular Conformer Generation with SO(3)-Averaged Flow Matching and Reflow
ICML (2025)
Adaptive Flow Matching for Resolving Small-Scale Physics
ICML (2025)
Proteina: Scaling Flow-based Protein Structure Generative Models
ICLR 2025 (oral, top 1.8%)
ProtComposer: Compositional Protein Structure Generation with 3D Ellipsoids
ICLR 2025 (oral, top 1.8%)
Energy-Based Diffusion Language Models for Text Generation
ICLR 2025
Truncated Consistency Models
ICLR 2025
2024
2023
LatentDock: Protein-Protein Docking with Latent Diffusion
Machine Learning for Structural Biology Workshop (NeurIPS 2023)
Bending and Binding: Predicting Protein Flexibility upon Ligand Interaction using Diffusion Models
Generative AI and Biology Workshop (NeurIPS 2023)
Compositional Score modeling for Simulation-based Inference
ICML 2023
Langevin Diffusion Variational Inference
AISTATS 2023
2022
Variational Inference with Locally Enhanced Bounds for Hierarchical Models
ICML 2022 (spotlight)
2021
2020
Empirical Evaluation of Biased Methods for Alpha Divergence Minimization
AABI 2020 (contributed talk)
Approximation Based Variance Reduction for Reparameterization Gradients
NeurIPS 2020
A Rule for Gradient Estimator Selection, with an Application to Variational Inference
AISTATS 2020
2018
2015
A little background
I obtained my PhD in Computer Science from UMass Amherst in 2023, supervised by Justin Domke.
During my PhD, I was grateful to intern at VantAI (New York, summer 2023) with Luca Naef and Michael Bronstein; Google DeepMind (London, summer 2022) with Andriy Mnih and George Papamakarios; Microsoft Research (Cambridge UK, summer 2021) with Cheng Zhang, Emre Kiciman and Miltos Allamanis; and Amazon AWS (New York, summer 2018) with Bing Xiang and Ramesh Nallapati.
Before my PhD, I also did some work on planning. Outside research, I try to play tennis.
Service & teaching
Conference reviewing
- NeurIPS
- 2019, 2020, 2021, 2022, 2025, 2026
- ICML
- 2020, 2025, 2026
- ICLR
- 2021, 2022, 2025, 2026, 2027
- AISTATS
- 2021
Organizing & journal reviewing
Publication chair, AISTATS 2023
Reviewer for JMLR and TMLR
Teaching assistant
- Probabilistic graphical models
- Machine learning
- Introduction to data structures
- Introduction to natural language processing
- Human–computer interaction