Python Developer for Podcast Transcription + AI Pipeline (Google Drive + Notion)

Project Overview I'm a CPA specializing in dental practice valuations. I want to build an automated system that monitors industry podcasts, transcribes new episodes daily, uses an LLM to extract substantive content, and delivers it to me in a searchable, organized format. After the forward pipeline is running, I'll also want to backfill historical episodes (potentially thousands). What the System Needs to Do Monitor approximately 24 dental industry podcast RSS feeds daily for new episodes. When a new episode appears, download the audio and transcribe it using a service like AssemblyAI or Whisper (open to your recommendation). Save the full transcript as a Google Doc in an organized Google Drive folder structure. Use an LLM (Claude or GPT-4) to extract substantive content — practice valuation insights, financial benchmarks, operational metrics, deal multiples, regional trends, specialty-specific information — while ignoring banter, ads, and pleasantries. Push the extracted nuggets into a Notion workspace as structured database entries, each linked back to the source Google Doc transcript. Generate a weekly email briefing summarizing the week's nuggets organized by theme. Architecture I'm Expecting Full transcripts live permanently in Google Drive as Google Docs, organized by podcast and date. Notion holds the structured metadata, extracted nuggets, and serves as the interface — but every Notion entry links back to its full transcript in Google Drive. This keeps my underlying data portable and not locked into Notion. Notion Setup Required Podcasts database (master list of monitored feeds with on/off toggles) Episodes database (one row per episode with metadata, link to Google Doc transcript, summary, and tags) Nuggets database (each extracted insight as its own entry, tagged by theme, dental specialty, and region, linked to source episode) Pre-built views: weekly briefing, by specialty, by theme, valuation-relevant only Tech Stack I'm Expecting Python, feedparser for RSS, transcription API (your recommendation), LLM API (Anthropic Claude or OpenAI), Google Drive/Docs API, Notion API, and scheduled execution (GitHub Actions, cloud function, or similar — open to your recommendation). Deliverables Working pipeline running on a schedule. All code in a GitHub repository that I own. All API keys and accounts created in my name. Notion workspace fully set up with databases and views. A short Loom video walking me through the system end to end. Documentation Requirements Written documentation that a non-developer can follow for day-to-day use, covering how to add or remove podcasts, how to modify the extraction prompt, and how to troubleshoot common issues. In addition, technical documentation sufficient for another competent developer to take over the system if you become unavailable. This should include system architecture, data flow, API integrations, scheduled jobs, error handling, environment configuration, and any non-obvious design decisions. Independent Third-Party Review and Final Payment Contingency Because I am not technical, I have engaged an independent third-party developer to perform quality assurance reviews throughout this project. The reviewer will: Confirm that the documentation meets professional standards and is sufficient for another developer to take over the system without your involvement. Verify that every component described in the documentation is actually built, deployed, and functioning as described. Identify any gaps, missing pieces, or quality concerns. Final payment of 20% of the total project budget will be held in escrow and released only after the third-party reviewer confirms the system and documentation meet professional standards. If the reviewer identifies deficiencies, you will be expected to address them as part of the project's quality assurance process before final payment is released. Note on Review Cadence The independent third-party reviewer will be engaged throughout the project, not just at the end. They will review your work at the completion of Phase 1, do a mid-build check-in during Phase 2, and conduct the final review before final payment. This is intended to catch concerns early when they are easiest to address, not to micromanage your work. The reviewer's role is bounded — they verify deliverables match what was promised and flag legitimate concerns, not redesign the system or impose stylistic preferences. If this structure doesn't work for you, this isn't the right project for you. Please factor all of this into your proposal and pricing. The escrow structure and review cadence are non-negotiable, and I'm being upfront about it so there are no surprises later. Post-Launch Support After delivery and the Loom walkthrough, I'm asking for the following support commitment as part of the engagement: First month after deployment: at least one hour per week of availability for troubleshooting, questions, or additional training as needed. Months 2 through 6 after deployment: availability on an as-needed basis for troubleshooting issues that arise and for additional training that we mutually agree to. We can structure this as a small monthly retainer or hourly as-needed — open to your preference. Please factor this into your proposal. Project Structure I want to do this in phases. Phase 1 is a paid trial of 5-10 hours building the monitoring and transcription piece for 2-3 podcasts so we can verify fit before committing to the full scope. Phase 2 expands to all podcasts and adds extraction, Notion integration, and briefings. Phase 3, as a separate engagement after the forward pipeline is dialed in, handles the historical archive backfill. Timeline Phase 1: about 1 week. Phases 1 and 2 combined: 2-3 weeks. In Your Application, Please Answer These Questions Describe a similar integration project you've built — ideally something involving multiple APIs, scheduled execution, and either Google Workspace or Notion. Which transcription service would you recommend for this use case and why? I'd particularly like your thoughts on speaker diarization since these podcasts have multiple hosts and guests. Have you worked with the Notion API before? Briefly describe a project where you used it. Please confirm you understand and accept the following: (a) 20% of the total project budget will be held in escrow until an independent third-party developer reviews the documentation and verifies the system meets professional standards; (b) the reviewer will be engaged at multiple milestones throughout the project, not just at the end; (c) you will address any deficiencies the reviewer identifies before final payment is released; and (d) the post-launch support commitment described above. Any concerns or questions about the scope as described? Applications that don't address all five questions specifically will not be considered. Budget Open to discussing based on your experience and approach. Please include your hourly rate, a rough estimate of total hours for Phases 1 and 2, how you'd like to structure the post-launch support pricing, and confirmation that you can accommodate the final-payment escrow structure. Apply tot his job Apply To this Job