Full Stack Developer Needed

We’re a small B2B agency building an internal lead sourcing and enrichment platform, and we need an experienced Full Stack Developer to help us fix and improve an existing module rather than build everything from scratch. The current system pulls company and contact data from multiple sources, cleans and standardizes it, enriches records using AI, scores leads against predefined criteria, and sends qualified leads into our CRM. The core workflow is already in place, but one of the existing modules is unreliable. When a source returns incomplete or inconsistent data, the enrichment/scoring flow can fail or produce incorrect results. We need someone who can understand the current codebase, identify the root cause, fix the module properly, and make the workflow more resilient. The stack currently includes Node.js, APIs, n8n workflows, a database, and Claude/OpenAI for enrichment and scoring. Depending on the existing architecture, you may also work on the frontend/dashboard used to review processed leads. We want AI to be used intelligently within the system, particularly for data normalization, enrichment, classification, and lead scoring. However, reliability matters more than simply adding AI calls. The workflow should handle missing fields, malformed API responses, duplicate records, timeouts, retries, and failed enrichment requests without breaking the entire pipeline. A key issue we want addressed is the current scoring module. Some incoming records have missing company attributes, and the existing logic does not handle these cases consistently. This can result in valid leads being incorrectly scored or rejected. We want this logic reviewed, corrected, and made fault-tolerant. We care about clean architecture, structured output, proper validation, useful logs, retry/error handling, and code that another developer can understand later. This is initially a focused project around fixing and improving the existing module, but if the collaboration goes well, we have ongoing Full Stack Development and AI automation work. Please do not send a generic proposal. At the end of your proposal, answer only the following: Share 1–2 relevant project links. Briefly explain your approach to identifying and fixing the existing module issue. Share any suggestion you have for improving the reliability or architecture of this workflow. Generic proposals will not be considered.