Machine Learning Engineer

Aether Biomachines•Menlo Park, CA

3h•$160,000 - $225,000

About The Position

At Aether, we are building a full-stack machine learning system to predict the outcomes of high-throughput laboratory experiments. Our goal is to transform noisy, real-world biological data into reliable, auditable, and predictive systems that accelerate scientific discovery. As a Senior Machine Learning Engineer, manage and improve the end-to-end ML and data pipeline that connects automated laboratory systems to production-grade models. This is not a model-tuning role. You will design the infrastructure, signal processing systems, and model architectures that turn experimental chaos into structured insight. You’ll work at the intersection of software engineering, applied ML research, and computational science - partnering closely with lab engineers and scientists to ensure our systems are robust, interpretable, and scientifically grounded.

Requirements

Python expertise (PyTorch, NumPy, Pandas, multiprocessing, Plotly/Matplotlib)
Experience developing ML systems in Linux and AWS environments
Strong understanding of Git-based workflows
Experience with database design (Django, SQL)
Deep grounding in algebra, statistics, and signal processing
Strong experience training, testing, and evaluating neural network models
Neural Networks, Transformers, Convolutional Networks, Graph Neural Networks
Model training, evaluation frameworks, and performance comparison
Comfort working in research-oriented, hypothesis-driven ML environments

Nice To Haves

PyTorch Lightning
Experience implementing or running bio-AI models (e.g., ESM, ESM-Fold, AlphaFold, Boltz)
Molecular dynamics tools (ASE, GPAW, GROMACS)
Cheminformatics tools (RDKit, SMILES, SMARTS)
Sequence alignment tools (BLAST, MSA)
Background in chemistry or physics (thermodynamics, kinetics, theoretical modeling)
Experience with LLM pretraining/fine-tuning, diffusion models, or reinforcement learning

Responsibilities

Design and maintain a full-stack ML system that predicts outcomes from high-throughput experiments
Build data pipelines that ingest, validate, normalize, and version experimental data
Ensure full auditability and traceability across the data lifecycle
Develop and maintain database schemas and views that power intuitive internal tools and lab-facing GUIs
Capture experimental metadata in structured, reliable formats
Apply best practices in database design, versioning, and interconnectivity
Create clear and actionable visualizations for scientific stakeholders
Build signal processing pipelines for noisy, automated lab data
Develop methods to detect anomalies and quarantine untrustworthy data
Recover signal from noisy inputs using statistical and algorithmic techniques
Leverage experimental controls to validate and cross-check outputs
Design, train, and improve neural network models to predict experimental outcomes
Evaluate and compare model architectures (Transformers, CNNs, Graph Networks, etc.)
Conduct scientific studies to validate modeling hypotheses
Improve predictive accuracy through both modeling innovation and signal engineering
Apply principles such as neural scaling laws to guide model development