Betway Global ir patērētājiem paredzēts tiešsaistes sporta likmju un iGaming zīmols, ko pārvalda Super Group (SGHC). Tā produkti ietver sporta likmju veikšanu un tiešsaistes kazino pieredzi, ko atbalsta reģionālā darbība, piemēram, Betway Africa. Zīmolam ir partnerības ar sporta komandām un pasākumiem, tostarp Arsenal, Manchester City, Springboks un SA20. Betway Global darbība ietver arī drošākas azartspēles veicinošas iniciatīvas, ko īsteno ar Super Group atbildīgas spēles programmas SuperPlan palīdzību, kā arī darbu ar kopienu un sponsorēšanu.
Senior AI Engineer
Build and operate production AI systems for an online gaming company, spanning Azure machine learning, retrieval-augmented generation, and agentic applications. Deliver secure, monitored solutions and move them from proof of concept to measurable business outcomes.
Apraksts
- Turn business needs into AI and machine learning solutions with defined success measures.
- Develop and tune models for classification, forecasting, recommendations, anomaly detection, and other applications.
- Prepare data through feature engineering, cleaning, and exploratory analysis across structured and unstructured sources.
- Build retrieval-augmented generation systems with embeddings, vector or hybrid search, reranking, and grounded responses.
- Connect and coordinate large language models using Azure OpenAI and Azure AI Foundry.
- Use prompt design, structured outputs, tool calling, fine-tuning, or distillation when they fit the solution.
- Assess generative applications for quality, speed, and cost, while managing token and inference usage.
- Create agentic systems that plan and complete multi-step work through tools and APIs, with safeguards and human review.
- Build multi-agent workflows with Semantic Kernel, Azure AI Agent Service, AutoGen, or LangGraph.
- Design agent tools, memory, state handling, and paths for errors and escalation.
- Develop language processing features such as entity extraction, summarisation, sentiment analysis, and intent detection.
- Implement speech-to-text for both batch jobs and live transcription.
- Use document intelligence and OCR to extract information from forms and documents.
- Release models and AI services through APIs, containers, or managed endpoints with automated CI/CD.
- Set up experiment tracking, model registries, version control, and repeatable training pipelines.
- Track production performance, drift, quality, cost, and reliability.
- Investigate operational issues and support business applications.
- Apply responsible AI practices, including fairness checks, content safety, evaluation tools, and clear statements of limitations.
- Meet privacy, data protection, and regulatory obligations, including controls for personal data and data residency.
- Protect systems with managed identities, key management, network safeguards, and least-privilege access.
- Keep models, prompts, data sources, and decisions auditable.
- Coordinate with data engineering on pipelines, feature access, and data quality.
- Work with software engineers to embed AI capabilities in products and internal platforms.
- Communicate methods, trade-offs, and outcomes to technical and non-technical colleagues.
- Document architectures, evaluations, solutions, and operational runbooks.
- Take AI projects from proof of concept to production with measurable business results.
- Meet agreed accuracy and latency targets for models and generative applications.
- Keep deployed systems reliable, supportable, monitored, and aligned with responsible AI requirements.
- Create shared components, pipelines, and practices for broader team use.
Prasības
- 3–5 years of practical experience in AI, machine learning, or data science engineering, including production deployments.
- A Computer Science degree is preferred; related engineering, information systems, data science, statistics, or quantitative degrees are also considered.
- Relevant professional certification is preferred, such as Microsoft Azure AI Engineer Associate (AI-102) or Azure Data Scientist Associate (DP-100).
- Hands-on experience delivering solutions with Azure AI and machine learning services.
- Demonstrated delivery in several areas, such as machine learning, RAG, LLM applications, agentic AI, NLP, or speech-to-text.
- Strong Python and SQL skills, with an emphasis on clean, tested, maintainable code.
- Understanding of model evaluation and the ability to validate solutions before release.
- Able to diagnose data, model, and integration problems effectively.
- Able to work independently and collaborate across functions.
- Willingness to keep learning.
- Experience with Microsoft Azure, Azure Machine Learning, Azure AI Foundry, and Azure OpenAI Service.
- Experience with Azure AI Search and Azure AI Services.
- Experience with Azure Functions, Azure Container Apps, and AKS.
- Familiarity with Azure Key Vault, Microsoft Entra ID, managed identities, Azure Monitor, and Application Insights.
- Experience with Azure Data Lake Storage, Azure SQL, Cosmos DB, Azure Databricks, and Microsoft Fabric.
- Strong SQL skills and experience handling large volumes of structured and unstructured data.
- Python experience using pandas, scikit-learn, PyTorch or TensorFlow, and Hugging Face Transformers.
- Experience with LangChain, LlamaIndex, Semantic Kernel, or a comparable orchestration framework.
- Experience with MLflow and Prompt flow.
- Experience with Git, Azure DevOps, GitHub Actions, Docker, and infrastructure as code using Bicep or Terraform.
- Experience building REST APIs, commonly with FastAPI or an equivalent framework.
- Credit and criminal record screening is required.
- A qualification verification check is required.
- Available to work in a hybrid arrangement in Johannesburg, with office attendance Monday through Thursday.
Priekšrocības
- Hybrid schedule with office work Monday through Thursday and remote work on Friday.
- Opportunities for professional growth and career development.
- Consideration for the Talent Management and Talent Development Programme.
- Opportunities to learn about and adopt emerging AI technologies.
- Some flexibility may be needed to support production releases or resolve incidents.