About the role

Join Stay22 as a Senior ML/DL Developer in the Neuro Squad. Design and optimize AI solutions, enhancing affiliate platform performance and user experience.

Responsibilities

Own the backend logic and model behavior for Roam and Spark, ensuring high accuracy and reliability.
Continuously improve models for ranking quality, latency, and revenue efficiency using advanced statistical analysis and deep learning.
Develop high-performance models for forecasting and predicting user intent from complex tabular and booking data.
Design and maintain the internal LLM client and gateway layers to provide unified access to language models across the organization.
Design and operate the full ML lifecycle: data preparation, training pipelines, model registries, and real-time inference services.
Manage caching layers, inference servers, and performance monitoring to ensure low latency for our high-volume engines.
Deploy robust, maintainable, production-ready solutions aligned with CI/CD and security standards (in collaboration with the Forge Squad).
Collaborate with the Data Squad to co-design training datasets and feature pipelines, ensuring smooth integration with the data warehouse and analytics systems.
Expose ML capabilities as robust APIs and services, allowing aligned squads (e.g., Hub, EmbedX) to integrate them easily without handling ML complexity.
Act as a mentor and technical advisor, supporting teams in prototyping ML features and adopting AI best practices.

Fluent English required for daily communication with international colleagues and partners.
Master's degree (M.Sc.) in Computer Science, Machine Learning, Data Mining, Statistics, or a related technical field.
6+ years of experience in machine learning or data science.
At least 2 years of experience in software engineering, including writing backend or production-ready API code.
Excellent command of Python and deep learning libraries such as PyTorch or TensorFlow, as well as scikit-learn.
Strong understanding of modern ML techniques, including CNNs, LSTM/RNNs, and gradient boosting models (XGBoost, LightGBM) applied to tabular data.
Proven ability to implement and optimize LLMs, including Transformers, custom embeddings, and building LLM gateways or clients.
Experience with Google Vertex AI (preferred) or Amazon SageMaker for training and managing large-scale models.
Extensive experience designing inference systems and real-time APIs (beyond batch processing).
Experience with MLOps tools (e.g., MLflow, Kubeflow, TFX) and containerization (Docker, Kubernetes).
Excellent SQL skills and experience collaborating on feature stores or ETL definitions.
Ability to design reusable, robust, and easy-to-consume systems for other teams.
Strong interest in translating complex ML metrics into concrete business value (better routing, higher conversion rates).