Senior Full-Stack Developer for Twilio + OpenAI + Google Calendar MVP

Description: We are building an MVP for an AI assisted missed inquiry recovery and intake automation tool for service businesses that depend on inbound customer inquiries. The product helps businesses respond quickly when an inbound inquiry is missed, collect basic intake information through an AI assisted conversation, schedule a callback appointment, send reminders, and provide a simple dashboard to review leads, conversations, appointments, and outcomes. Potential use cases include law firms, elective medical practices, home services, professional services, and other businesses where missed inbound inquiries can mean lost high value leads. Core workflow: Missed inquiry to automated follow up to AI assisted intake to appointment booking to reminders to dashboard and reporting. This is not intended to be a full CRM, case management system, EHR, or marketing automation platform. It is a focused pilot MVP for one initial business. The MVP should prioritize working workflows, reliability, clean architecture, and speed of implementation over visual polish. This is initially a one business pilot. However, the app should be built with a multi business ready backend architecture. We do not need full SaaS onboarding, billing, white label admin, or complex user permissions yet, but the core data model should support adding more businesses later without a rewrite. Important integration requirement: The tool should sit on top of the business's existing inquiry handling setup wherever possible. The pilot should not require the business to replace its existing provider or change its public facing customer inquiry flow unless explicitly approved. Preferred implementation is to receive missed inquiry, after hours, abandoned inquiry, or similar events from an existing provider, call tracking setup, or communications API through webhooks or API events, then trigger the automated follow up workflow. Normal inbound inquiries should continue routing to the business's existing staff, receptionist, call center, or current handling process. Preferred stack: Open to recommendations, but preferred stack is: Next.js Node.js TypeScript Supabase Postgres OpenAI API Google Calendar API Vercel Background jobs or scheduled reminders Simple authentication such as Supabase Auth, Clerk, or similar Core MVP features: The MVP should include: Inbound inquiry event handling Missed inquiry detection Automated follow up workflow Lead or contact record creation Conversation logging Opt out handling English and Spanish AI assisted intake flow Configurable intake questions Structured intake state management Safe response guardrails so the AI does not make promises, give regulated professional advice, or overstep Human review and escalation flags Calendar availability lookup Appointment booking Customer calendar invite Confirmation workflow Reminder workflow Business notification with intake summary Simple lead dashboard Lead detail page with conversation transcript Manual status and outcome tracking Basic ROI or recovered lead reporting Basic activity and event logging Simple settings for the pilot business AI and LLM requirements: The AI should be used for: Language detection Natural language parsing Classifying inquiry type Continuing the conversation in the user's language Collecting basic structured intake information Summarizing the conversation for the business Flagging urgent, confusing, hostile, sensitive, or unusual conversations Creating a concise internal summary in English The AI should not: Give legal, medical, financial, or other regulated professional advice Promise a result Claim the business will accept the customer, client, or patient Estimate outcomes, pricing, value, or eligibility unless explicitly configured Create a professional relationship on behalf of the business Ask for unnecessary sensitive information Make guarantees Negotiate anything on behalf of the business The system should be designed with conservative guardrails and clear handoff to a human when needed. Example intake behavior: The system should ask short, simple questions that help the business decide how to follow up. Examples of configurable intake questions: What are you reaching out about? When did this happen or when are you hoping to be seen? Is this the best way for the team to reach you? What time works best for a callback? Would you like to continue in English or Spanish? The exact intake questions should be configurable by business type later. For the pilot, a simple default flow is acceptable. Calendar booking requirements: The MVP should support booking a callback appointment after intake is complete. The business should be able to configure basic scheduling rules such as: Business hours Appointment length Buffer time Time zone Same day appointment rules Reminder timing Internal notification recipient Optional escalation recipient The system should be able to: Look up available appointment times Offer a few appointment options Book the selected appointment Create the internal calendar event Send the customer a calendar invite Send confirmation and reminder messages Update the lead record with appointment details Advanced scheduling features can be deferred if needed, but the pilot should reliably book and log appointments. Dashboard requirements: The MVP should include a simple dashboard for the business. Lead list should show: Lead name if collected Status Language Inquiry type Appointment time Last activity timestamp Outcome Human review flag Lead detail page should show: Lead information Inbound event history Conversation transcript AI intake summary Appointment details Language Status Outcome Notes Manual status update Human review flag Suggested lead statuses: New missed inquiry Follow up sent Conversation active Intake complete Needs human review Appointment offered Appointment booked Reminder sent Completed Retained or converted Not a fit No show Reschedule needed Unresponsive Opted out ROI and reporting requirements: The MVP should include lightweight reporting. Business should be able to enter basic assumptions such as: Average customer value Average conversion rate Monthly subscription cost or estimated software cost Dashboard should show: Missed inquiries captured Follow ups sent Response rate Intake completion rate Appointments booked Appointments completed Converted customers or retained clients Estimated recovered revenue Estimated ROI The reporting does not need to be advanced attribution. It just needs to help the business understand whether the system is recovering leads that would otherwise have been lost. Multi business ready architecture: This is a one business pilot, but the backend should be structured for future expansion. Every core record should be scoped to business_id, account_id, or similar, including: Businesses or accounts Business settings Leads or contacts Inbound events Conversations Messages Appointments Activity logs Reports We do not need: Self serve onboarding Billing White label admin Complex user permissions Full CRM functionality Deep industry specific integrations Advanced attribution Complex marketing automation But adding a second business later should not require a rewrite. Expected milestones: Please provide your own recommended milestone plan, but we expect something like this: Milestone 1: Architecture, project setup, database, authentication, deployment, and initial data model. Milestone 2: Inbound event handling, missed inquiry detection, automated follow up workflow, conversation logging, and opt out handling. Milestone 3: OpenAI integration, language detection, AI assisted intake flow, structured conversation state, summaries, and human review flags. Milestone 4: Calendar availability lookup, appointment booking, confirmation workflow, reminder workflow, and appointment updates. Milestone 5: Dashboard, lead list, lead detail page, transcript view, manual status updates, notes, and outcome tracking. Milestone 6: ROI reporting, business notifications, QA, bug fixes, production readiness, and handoff documentation. Ideal candidate: You should have experience with: API integrations Webhook based workflows OpenAI API or LLM workflows Structured multi step conversation state Calendar API integrations Next.js Node.js TypeScript Supabase Postgres Background jobs Scheduled reminders API reliability Logging Error handling Simple dashboard development Bonus if you have built: Appointment booking tools AI intake or chat workflows Lead management dashboards Workflow automation products Products for high value inbound lead businesses Communication API integrations Customer follow up automation Multi tenant SaaS architecture Budget and structure: We expect this MVP to be roughly 350 to 500 hours depending on implementation details. Preferred hourly range is 35 to 55 USD per hour for a senior developer. We are open to hourly or fixed price milestone proposals, but we want candidates to provide a clear milestone breakdown and estimated hours. This is initially a short term MVP contract, but if the pilot is successful, there may be ongoing maintenance, support, future feature development, or potentially a longer term role. Timeline: Target timeline is 6 to 8 weeks. Speed matters, but reliability matters more. This product will handle real inbound leads, so broken workflows, missed follow ups, failed bookings, or failed reminders are serious issues. Paid technical trial: Shortlisted candidates may be asked to complete a small paid technical trial before the full MVP contract is awarded. Example trial: Build a small prototype that receives an inbound event, stores it in Supabase, calls OpenAI to classify language, intake stage, and urgency, and returns a safe next step response. This will be paid. We are not asking for free work. To apply: Please answer these questions: 1. Have you built production API or webhook based workflow tools before? Please describe. 2. Have you built OpenAI or LLM workflows with structured state or multi step conversations before? 3. Have you integrated calendar availability and event booking before? 4. How would you structure this one business pilot so it can support more businesses later? 5. What do you see as the biggest technical risks in this project? 6. What stack would you recommend and why? 7. What timeline and estimated hours would you propose? 8. How would you handle retries, logging, and failed workflow events? 9. How would you keep the AI from overstepping or giving regulated professional advice? Please describe relevant prior work in your proposal.