Senior Machine Learning Engineer

Hum's mission is to learn how the Earth changes by building a “world model for the entire world”: one that represents the dynamics of Earth processes, forecasts what a place looks like next, and makes that understanding useful. Founded at the University of Waterloo and backed by top investors like HF0, Climactic, Onto Ventures and angels from Firebase and OpenTable, we are a small team working on some of the largest unsolved problems. As our Senior Machine Learning Engineer, you'll own the software and infrastructure that trains and serves our models, so our research scientists can stay focused on the models themselves.

Responsibilities

Own the ML platform: build and run the systems for experiment tracking, distributed training at scale, and serving inference, and keep them fast, reliable, and reproducible; Collaborate on the interfaces: work with engineers and scientists to define and maintain the boundaries between data and training, data and inference, and inference and the agent; Enable multimodal training: build the data and model plumbing that lets us train on many sensors, observation types, and text in one model; Go from small to large: help scientists prove ideas in fast, cheap experiments, then scale the winners to petabyte-scale training runs; Make evaluation real: implement and improve benchmarks and evaluation frameworks, both internal and user-facing; Watch the meter: measure and optimize the cost and wall-clock time of experiments, training, and inference; Support the agent: work with the product team on how the agent calls the model and consumes its outputs, and build the integration

Requirements

Bachelor's degree in computer science, engineering, or a related field, or equivalent experience; 5+ years of relevant industry experience, including hands-on machine learning engineering; Proficiency building distributed, multi-node training pipelines in PyTorch and Lightning; Experience with AWS or another cloud platform, particularly the tooling for training and serving medium-to-large models; Strong software engineering fundamentals: you write production-quality Python and care about reliability, observability, and cost; Fluent with modern agentic coding tools (e.g. Claude Code, Cursor, Codex) and current on how to use them well

Nice to have

Experience training large diffusion or transformer models, ideally on video or time-series data; Experience with multimodal or spatiotemporal data, geospatial data, or geoscience problems; Experience serving models in production and integrating them with downstream applications or agents; Startup experience, comfortable with a small, fast-moving team

Benefits

Competitive salary; meaningful early-stage equity; health, dental and vision coverage