About the Role
The Machine Learning Engineer takes models from prototype to production and keeps them working there. Where our AI Systems Engineer role focuses on applying AI to product problems, this role goes deeper into the modelling itself: training and evaluation, feature pipelines, inference performance, and the monitoring that catches a model degrading quietly. The work is applied throughout — the goal is a measurable product improvement in production, not a result on a benchmark.
What You'll Do
- Build, train, and evaluate models for real product problems.
- Design evaluation that reflects product outcomes rather than convenient metrics.
- Build feature and training pipelines that are reproducible.
- Deploy models to production and own their latency, cost, and reliability.
- Monitor for drift and degradation, and retrain deliberately.
- Run experiments and A/B tests that support honest conclusions.
- Work with engineers and product to decide where a model is the right answer — and where it is not.
What We Look For
- Experience taking machine learning models into production and keeping them there.
- Strong Python and familiarity with the modern ML stack.
- Rigorous approach to evaluation, baselines, and experiment design.
- Understanding of inference cost, latency, and serving trade-offs.
- Software engineering discipline: tests, versioning, reproducibility.
- Ability to explain model behaviour and limitations to non-specialists.
- Honesty about uncertainty — knowing when a result is not yet real.
Nice to Have
- Experience with recommendation, ranking, matchmaking, or personalisation systems.
- LLM fine-tuning, evaluation, or retrieval-augmented systems in production.
- Experience with fraud, abuse, or anomaly detection.
- Experience with simulation or reinforcement learning in games.
About VistralNova
VistralNova is a technology company building developer tools, game systems, AI-powered experiences, and digital infrastructure for modern interactive products. Our work brings together software engineering, game design, artificial intelligence, and asset-ready infrastructure to help teams create products that are practical, scalable, and easier to trust.
We are evolving toward a flexible technical foundation aligned with Polkadot and PVM-based architecture. Our team values clean architecture, practical execution, strong engineering standards, and long-term product quality.
