Software Developer (AI and Data Platform)

Job description

We are looking for a Software Developer to support and enhance Agent Spendster, an AI-powered spend compliance platform that automates procurement and invoice validation across the procure-to-pay lifecycle. The role involves building scalable data pipelines, integrating enterprise systems, productionizing AI solutions, and ensuring reliable, secure, and observable services on Azure cloud. **Roles \& Responsibilities:** * Develop and maintain scalable AI, data integration, and automation pipelines. * Enhance Azure Durable Functions workflows, orchestration, error handling, and performance optimization. * Build and support integrations across Snowflake, SAP, Azure services, and enterprise data sources. * Productionize AI solutions and resolve platform, pipeline, and data quality issues. * Build evaluation and monitoring frameworks for AI/ML workloads. * Implement observability, logging, tracing, dashboards, and operational monitoring. * Improve automated testing, CI/CD pipelines, authentication, and platform reliability. * Collaborate with Data Scientists to operationalize AI models and document intelligence solutions. **Technical Requirement:** * Strong Python development experience with Python 3.x, Pydantic, Pytest, Type Hints, and API development. * Experience building cloud-native solutions on Microsoft Azure, including Azure Functions, Durable Functions, Azure OpenAI, Document Intelligence, Cosmos DB, Blob Storage, and Key Vault. * Strong data engineering and integration skills with SQL, Snowflake, ETL/ELT pipelines, and enterprise data platforms. * Knowledge of CI/CD, Git, GitLab, DevOps practices, observability, monitoring, and logging frameworks. * Experience with OpenTelemetry, Application Insights, Grafana, and production monitoring tools. * Understanding of secure authentication mechanisms including Managed Identity, OAuth2, and Workload Identity Federation. * Exposure to LLM applications, Azure OpenAI, OCR/Document Processing, SAP integrations, MLflow, and Azure ML is preferred.