Python/AI Developer: Workflow Automations & Data Analyses

About the project We run account management for e-commerce/dropshipping clients. A large part of the work is recurring workflows: order checks, delay and compensation handling, disputes, stock control, and reporting. We're automating these. You'll be responsible for the workflow automations and data analyses end-to-end — a mix of regular automation and AI-driven actions. This is one of two parallel tracks. A separate developer is building a WhatsApp workspace ("skin") for our account managers, and your automations will ultimately surface inside that workspace. The two tracks are built separately but must integrate cleanly, so well-defined interfaces and good coordination matter. Who you are - AI-first developer, strong in both Python and AI, with real coding skills under the hood. - You build microservice-based, not monolithic. Our hard requirement: each automation and service is isolated so that if one breaks, the rest keeps running. You design with clean separation and replaceable parts. - Comfortable with a mix of regular automation and AI actions — you know when a rule-based approach is right and when to reach for AI, and you can build the AI-judgment parts well (e.g. "is this really out of stock or just delayed?", summarizing, drafting). - Able to work with APIs and handle real data volumes efficiently (order datasets can get large). - Responsible for work end-to-end: architecture, build, testing, delivery. - Good communicator; we work closely (WhatsApp + 1/2 weekly calls), with detailed workflow docs provided. What you'll build - Workflow automations: rule-based work like filtering, status checks, reminders, and reporting, driven off our platform API. - AI actions and data analyses: the judgment-based tasks and analyses over large datasets. - A clean, maintainable, microservice-based setup that plugs into our wider system — your automations will connect to and surface inside the WhatsApp workspace built by the other developer, so clean integration points are part of the job. Good to know The AI needs to run hosted (via API) to start, built behind clean adapters so some parts can move to local/self-hosted models later where it makes sense. Experience with both hosted AI APIs and running/optimizing local/open-source models is a real plus. How we work - Task-by-task roadmap, detailed workflow docs, no sprints — you build, we review per delivery via demos and testing. - You report to our project lead, who coordinates everything (and coordinates between the two tracks). To apply tell us briefly: (1) where you've combined rule-based automation with AI, and how you decided the split; (2) an example of a microservice-based system you built end-to-end, ideally one that had to integrate with another system or team's work; (3) your experience with hosted AI APIs and/or self-hosted models. Short and concrete beats long and generic.