AI/ML engineer for content moderation

Besedo; United States; Remote; USD 140.000 - 190.000; Vollzeit

Benefits dieser Stelle

Удаленная работа 100%

Гибкий график

Zusammenfassung

Besedo Global Services AB ищет старшего инженера по машинному обучению, который возьмет на себя эксплуатацию ML‑моделей в продакшене: развёртывание, мониторинг и оптимизация инференса (ONNX/TensorRT).

Кандидат будет участвовать в разработке конвейеров обучения и переналадки моделей, обеспечивать воспроизводимость, настройку CI/CD и масштабируемость сервисов, тесно сотрудничая с DevOps и командами аналитики.

Qualifikationen

• Опыт разворачивания ML-моделей в продакшене и мониторинга производительности

Aufgaben

• Владение жизненным циклом ML-систем: обучение, упаковка, развёртывание, мониторинг и контролируемые релизы
• Повышение эффективности инференса: ONNX/TensorRT, батчинг, настройка параметров
• Работа с DevOps/SRE для обеспечения эксплуатационной устойчивости: масштабирование, наблюдаемость, оповещения
• Контроль версионности моделей, CI/CD и воспроизводимость в продакшене
• Контроль качества моделей и валидация в условиях меняющихся данных и бизнес‑потребностей

Kenntnisse

Python

PyTorch

MLOps

CI/CD for ML

ONNX/TensorRT

Docker

Ausbildung

Master’s/PhD in CS/Math or related

Tools

GitLab CI

Kubernetes

SageMaker

Jobbeschreibung

Описание:

At Besedo we are looking for a Senior Machine Learning Engineer to own how our ML models run in production, from deployment and monitoring to inference optimization (ONNX/TensorRT). Besedo is a tech company delivering Trust & Safety solutions to online marketplaces worldwide. Implio is our all-in-one SaaS moderation platform that combines state-of-the-art AI (deep learning, NLP, computer vision) with a powerful user interface for human moderation and powerful analytics.

Задачи:

Handle the production lifecycle of ML systems: training, packaging, deployment, monitoring and controlled rollouts. Design, build, and continuously improve training and retraining pipelines (data ingestion, labeling/QA loops, feature generation, experiment tracking, reproducibility). Build and improve high-performance inference services and deployment patterns. Drive inference optimization (ONNX/TensorRT, quantization, batching, GPU/CPU tuning) with measurable latency/cost improvements. Partner with DevOps/SRE to ensure operational excellence: scaling, reliability, incident readiness, observability, and alerting. Own model versioning, CI/CD, and release workflows ensuring reproducibility and safe delivery to production. Strengthen data & model monitoring (performance, drift, quality) and implement feedback loops for continuous improvement. Communicate results and trade-offs to stakeholders, turning technical outcomes into business value. Ensure model quality through rigorous evaluation and production-grade validation aligned with evolving data and business needs.

Требования:

• Master’s, PhD, or equivalent experience in Computer Science, Engineering, Data Science (or related), with strong foundations in mathematics, statistics, and ML.
• 6+ years (or equivalent) experience across ML engineering and MLOps / production ML, with a proven track record delivering high-impact systems in production.
• Demonstrated expertise deploying models to production: scalable serving patterns, model packaging, inference optimization, observability, and safe rollout strategies.
• Strong experience with model performance and efficiency optimization: latency/throughput tuning, batching, quantization, ONNX/TensorRT, GPU/CPU optimization, memory and cost tuning.
• Strong proficiency in Python.
• Hands‑on experience with ML frameworks such as PyTorch, scikit-learn, and optimization/runtime stacks like ONNX, TensorRT.
• Strong grasp of CI/CD for ML (e.g., GitLab CI), containerization (Docker), and reproducible workflows.
• Experience with monitoring and reliability: metrics/logging/tracing, model/data drift, alerting, SLOs, incident response practices.
• Будет плюсом: Kubernetes, AWS, FastAPI (and/or similar Python serving frameworks), SageMaker, service meshes, feature stores, model registries

Условия:

100% Remote FTE or B2B (Romania, Malta) B2B (EU)