Build an AI-Powered Lead-Scoring Tool (Research → Score → HubSpot) — Node/TypeScript + LLM

We're a small, growing company looking for a developer to build a simple internal tool that helps us quickly size up potential customers. The idea is straightforward. You give the tool a company name and website, and it gathers some useful public information, applies a scoring approach we've already designed, and saves the result into our CRM (HubSpot). We'll provide the full scoring logic and any access you need — you just build the tool around it. Roughly, the tool should: - Pull together basic public info about a company from its website and a few other public sources. - Run that through our scoring approach to produce a simple score and a clear recommendation. - Save the results into HubSpot. - Offer an easy way to run one company at a time, or a list from a spreadsheet. We're flexible on how it's built — propose whatever is reliable and easy to maintain. For context, we work mostly with Node and TypeScript, and the scoring step will use an LLM (the Anthropic Claude API) with our prompt. We'd like the scoring settings to be easy to tweak later without redeploying. A few things we care about: - Solid Node / TypeScript skills. - Experience using LLM APIs in real projects. - Comfort pulling data from public sources / light scraping. - HubSpot (or similar CRM) API experience is a plus. - Clean, well-documented code, and no making up data when something isn't available. Timeline is roughly a few weeks. There's a good chance of ongoing work if the first version goes well. If you're interested, tell us briefly about something similar you've built, how you'd approach pulling the data, and your availability over the next month.