- Project Role :
- Machine Learning Engineer
- Project Role Description :
- Take models from a notebook to something that runs reliably inside a customer's operation, and stays honest once real data starts arriving.
- Must have skills :
- Python, PyTorch, model deployment
- Good to have skills :
- MLOps tooling, time-series forecasting, ONNX
- Experience required: 3–6 years
- Educational Qualification :
- Bachelor's or Master's in a quantitative field, or equivalent practical experience
Summary:
Most of this role is the distance between a model that works on a held-out set and one that holds up in production, drift, latency budgets, inputs that arrive late or malformed, and the question of what the system does when it is not confident. How a model is evaluated and documented is part of the work here rather than paperwork after it.
Roles & Responsibilities:
1. Modelling
- Frame problems with the people who own them before choosing a method.
- Build, train and evaluate models in Python and PyTorch.
- Establish baselines that a simpler approach has to beat.
2. Deployment
- Package and serve models with clear latency and cost characteristics.
- Instrument inputs and outputs so drift is visible before it is a problem.
- Define fallback behaviour for low-confidence and out-of-distribution cases.
3. Governance
- Document data lineage, evaluation and known limitations as a matter of course.
- Work with the AI Governance Lead on risk assessment before a model ships.
