Senior Applied Deep Learning Research Engineer

We are hiring a Senior Applied Deep Learning Research Engineer to work on the model itself — the architecture, training, and evaluation of the Meshwork Virtual Machine.

What you will work on

Foundation Model Research

• Own research threads on the architecture of the Meshwork Virtual Machine — multimodal representation learning, representation of heterogeneous biological measurements, simulation of dynamics, and perturbation response prediction
• Design and run training runs and ablations, and draw conclusions that survive scrutiny
• Own evaluation that distinguishes real capability from leakage, batch artifacts, and wishful metrics
• Bring methods from the wider machine learning literature into a domain that has not yet applied them

Engineering and Infrastructure

• Build the training and evaluation infrastructure the research requires
• Develop data pipelines that turn heterogeneous experimental data into model-ready inputs
• Own reproducibility for your work — experiment tracking, seeds, environments, and results others can rebuild
• Make and defend practical decisions about compute, scale, and where efficiency actually matters

Collaboration and Scientific Contribution

• Work directly with the biological team to turn biological questions into model questions, and results back into experiments
• Learn enough of the underlying biology to ask good questions and to recognize when a result is biologically implausible
• Contribute to technical documentation, internal reports, and investor-facing scientific material
• Share work externally where it serves the platform — publications, preprints, and open source

Profile

Experience or demonstrable ability is required in each of the following. Candidates are assessed on the substance of their work rather than on formal credentials alone.

Requirements

• Command of modern deep learning — architectures, training dynamics, optimization, and evaluation
• Hands-on experience training and evaluating large models on real data, beyond fine-tuning existing checkpoints
• Proficiency in Python and PyTorch or JAX, with the engineering ability to build experimental infrastructure independently
• Judgment to own research direction rather than only execution — defining a problem, designing the experiment that tests it, and determining when a result is sound
• Ability to communicate clearly with scientists who are not machine learning specialists
• Capacity to acquire domain knowledge in an unfamiliar field, and the rigor to do so properly

Advantages

Domain experience in the life sciences is an advantage rather than a requirement; the necessary biological grounding is developed in role.

• Experience with transformers, graph neural networks, diffusion or generative models applied to structured scientific data
• Foundation model work — pre-training, adaptation, or rigorous evaluation
• Multimodal or cross-modal modeling, including work on modalities with mismatched scale and noise
• Research experience at a frontier AI lab or a leading AI-for-science organization
• A PhD in machine learning, computer science, or a related field
• Exposure to biological, biomedical, or other high-noise experimental data
• Open-source authorship or maintenance
• Experience in a startup or early-stage research environment