Forward Deployed Machine Learning Engineer
Confirmed on the employer's own hiring board on Oct 28, 2026. First seen by Alion on May 11, 2025. Stratum scores D on the Alion truth index.
We are looking for a Forward Deployed ML Engineer, with an interest in solving difficult real-world problems to join our Technical Services team.
This role will require extensive working with custom architectures on PyTorch. Previous experience working with PyTorch on complex Convolutional Neural Network, Graph Neural Network, and/or Transformers is expected. This role is equivalent to the Delta team Forward Deployed Engineer role at Palantir that we would recommend you read up on.
This is a remote-first position, with a preference for applicants based in Canada.
Technical Services Team Responsibilities
Conducting foundational research to design and implement more accurate AI resource modeling techniques that we can use for many different mining operations.
Applying and refining existing architecture to train resource models for a specific mine.
Communicating the quality, metrics, performance, and methodology of our models to non-ML technical external stakeholders.
Identifying new opportunities to improve mining operations (such as modelling other parameters) for a given client and getting their support in doing so.
Tracking model performance of deployed models over time and identifying ways to make existing models better.
Role Responsibilities
Adapt Stratum’s deep learning models to a given mine.
Develop and maintain high-quality machine learning code using Python.
Unlock ways to create better resource and metallurgical models with deep learning for one specific mine, or sometimes mines in general
60% of your time is dedicated to applying ML to one specific mine (applied ML), 40% is dedicated to applying ML to mines in general (foundational ML).
Take part in at least 2 mine visits a year and engage with our clients on the ground
Over time, grow into a senior engineer who can identify additional ways our technology can be leveraged with our existing clientele
Thrive in ambiguity with complex, large datasets and evolving client needs - able to shape these into clear, actionable and structured actions
Role Requirements
2+ years of industry machine learning experience.
Excellent proficiency in implementation of custom neural network architectures in Python (Pytorch).
The ideal candidate will have hands-on experience with various neural network types
Strong foundational knowledge of deep learning with an emphasis on transformers, convolutional neural networks, and/or graph neural networks.
Adept at working with a high degree of autonomy and initiative to solve complex problems.
Strong technical communication skills (oral + visual).
You will be responsible for conveying to other non-machine learning engineers in the mining industry why our models are good. You will work with geologists and business development roles to tell the story.
Nice to Have
Ability to speak another language with high proficiency.
Our Tech Stack
AI Stack: Python & Pytorch
Backend: Python (FastAPI) + Pydantic + DynamoDB
Infrastructure: AWS, Azure
About Stratum
We're Stratum, a mining software company with machine learning models as our core product. Our 3D maps predict how gold, silver, copper, etc. are distributed (and how much!) using only small amounts of data, unconventional data processing, and proprietary ML protocols. Our work directly affects how much money a mine is going to make next week/month/year while reducing waste/cost. We're supported by Founders Fund, Aramco, Builders VC, Y Combinator, and Ilya Sutskever, former Chief Scientist at OpenAI, who have recognized the potential of our industry-disrupting technology.
Our long-term vision is to build a massive AI engine capable of making every decision in a mining operation, down to moving individual rocks. If you’re an exceptional engineer interested to helping make this vision a reality we look forward to reviewing your application and working together.