[Hiring] Contract implementer (remote, 1099): install and support client sites + AI phone line on Next.js, Vercel, Supabase, Twilio, ElevenLabs, Cloudflare Workers

Jobs
1

Bridge Builders (Bridge Builders AI LLC, Florida) needs a contract implementer to stand up and support client installs of three products: a Business Website with the BridgeView dashboard, an AI Customer Assistant (a website chat assistant), and an AI Receptionist (an AI phone line). The engine is built and running at bridgebuildersai.com. You install it, client by client, against a written runbook, and keep it running.

What an install is:

  • Port the engine into a private GitHub repo for the client, layer their config and content, get typecheck and build clean.
  • Deploy prebuilt output through GitHub Actions to the client’s project in the Bridge Builders Vercel team; the project is never connected to Git.
  • Provision a per-client Supabase project, apply the migrations, set the Vercel environment from our env contract.
  • Verify the client’s Resend sending domain.
  • Guide the client through their own Twilio account: number, voice webhook, and a fallback TwiML Bin that dials their own phone.
  • Configure the client’s ElevenLabs agent in our workspace, apply the provider privacy policy (no stored audio), and read it back.
  • Deploy the Cloudflare Worker media gate.
  • Run the verification gates: build and render, the consent gate end to end, the Twilio failover test, secrets server-only.
  • Live call test, then the client walkthrough. Then ongoing support for your live clients, inside a monthly cap.

Stack: Next.js and TypeScript, npm workspaces, GitHub Actions, Vercel CLI, Supabase (Postgres, Auth, RLS, migrations), Resend, Twilio (Voice, TwiML, webhooks), ElevenLabs Agents, Cloudflare Workers and Wrangler, and OpenAI reached through a Bridge Builders gateway Worker. The LLM key behind the Receptionist is owner-held; you never touch it. We do not train the stack. We hand over our conventions: repo layout, env contract, provisioning flow, the receptionist consent boundary.

Pay (independent contractor, 1099):

  • Fixed fee per install: $300 Website + BridgeView, $300 AI Receptionist, $300 standalone AI Customer Assistant, $199 for the Assistant added to a site we built.
  • Fixed monthly share per live client for as long as they stay: $20 Website, $20 Assistant, $100 Receptionist.
  • The Receptionist share covers roughly 63 minutes of support per client per month; beyond that you bill $95/hr.
  • Never clawed back.

The honest framing: at 9 to 11 modeled hours, $300 is below the $35 to $50/hr market for this work. The recurring share is where it becomes worthwhile: by roughly month five a Receptionist client has paid you at our own modeled $85/hr labor value, and every month after is margin. This is a book of live clients, not a string of gigs.

Hiring step: a paid screening task, $150 to $250, about four hours: a real slice of an install against a written spec. Before any payment, a contractor agreement (confidentiality, IP assignment, no authority to bind Bridge Builders, the support cap) and a W-9. We contract for availability, not a schedule: 20 to 30 hours a week.

Reply with exactly three things:

  1. Links to things you have built and deployed (repos, live sites, Workers, agents).
  2. Your rate expectations relative to the fixed fees above.
  3. Your weekly availability, in hours.

Replies missing any of the three will not be read further. hello@bridgebuildersai.com

3 Likes
2

Hi, I saw your request: [Hiring] Contract implementer (remote, 1099): install and support client sites + AI phone line on Next.js, Ve…. Is this still open? Could you share the current steps and one example of the result you need? What budget and delivery date have you allocated? I would confirm a small fixed scope and price after reviewing the details.

3

Three things, in your order.

1. Built and deployed

  • github.com/koraynar/doc-extract-mcp and github.com/koraynar/trendyol-seller-mcp — two MCP servers, both published to the official MCP registry (registry.modelcontextprotocol.io/v0.1/servers?search=koraynar returns them) by a GitHub Actions workflow that authenticates with OIDC id-token: write, so the publish path holds no secret at all. That habit is what your env contract and owner-held LLM key actually need.
  • github.com/koraynar/t3n-approve-invoice — TypeScript. A policy-gated accounts-payable agent where the pay-to destination comes only from a local policy file and never from the incoming document, no matter what the document says. Same shape as your consent boundary: what the code must refuse to do is the specification.
  • koraynar.github.io/aninda-asistan/ — live single-page customer-intake assistant demo, two scenarios. It states on its own face that it is a simulation and stores nothing, which is the same disclosure discipline as your no-stored-audio policy readback.
  • koraynar.github.io/research-samples/ — second deployed Pages site.
  • Two Flutter apps live on Google Play, shipped through fastlane and store review: signed builds, per-app config, release gates that fail closed.

Straight about the gaps, because you will find them in the screening task anyway: I work in TypeScript, Python and Rust and I run CI/CD and API integrations daily, and Next.js, Vercel, GitHub Actions and Wrangler are familiar ground. I have not run Twilio Voice, ElevenLabs Agents or Supabase RLS in production. Against a written runbook those are config-and-verify work rather than invention, but I would rather you priced that in now than discovered it on a client install.

2. Rate

Your fixed fees as published — 300 / 300 / 300 / 199 — no premium on the early installs. I read the month-five math the same way you do: the recurring share is the part worth having, and it only pays if the installs I hand over stay quiet. Support past the cap at your 95/hr. 40/hr for anything outside the runbook. Screening task at your stated 150 to 250.

3. Availability

20 to 25 hours a week, sustained, not a burst. Istanbul, UTC+3 — my afternoon and evening cover the US morning and early afternoon Eastern.

Two things better said before you spend screening budget than after. I am not a US person, so the contractor tax form is W-8BEN rather than W-9 (Turkey, US treaty). And I do client handovers in writing plus a recorded screen walkthrough, not live calls — provisioning, the verification gates and the Twilio failover test are no problem, but if a live call with each client is a hard part of the handover, say so now and I will step aside rather than waste the trial on it.

4
  1. Built and deployed

I’ve built and deployed production web applications and automation systems using Next.js, PostgreSQL, GitHub Actions, APIs/webhooks, structured validation, automated testing, and release workflows.

Examples:

  • Meydano — production sports booking platform built with Next.js + PostgreSQL, including booking lifecycle logic, payments, role-based workflows, production testing and deployment.

  • SAMAP — production construction operations platform with Next.js/NestJS, database-backed workflows, authentication/permissions, automated testing and deployment.

  • n8n quotation automation demo:
    https://hvac-ai-demo.app.n8n.cloud/form/9ba77429-8dcc-4810-a30a-860d100ba31b

I’m comfortable working from a strict runbook, environment contract and verification checklist. I have not yet operated Twilio Voice or ElevenLabs Agents in production, so I’d rather be transparent about that upfront.

  1. Rate expectations

I’m comfortable with your published fixed fees for the initial installs and with the $150–250 paid screening task. For work outside the documented install/support scope, I’d confirm scope and rate before starting.

  1. Availability

I can allocate approximately 20 hours per week and can overlap with US working hours when needed.

One onboarding point to confirm before the screening task: I’m a non-US independent contractor, so I would need to use the appropriate non-US tax documentation (such as W-8BEN rather than W-9). Please confirm whether that works for your contractor setup.

Best,
Naser

5

I’ve deployed an AI automation system with GitHub Actions, PostgreSQL, API/webhook integrations and Cloudflare-backed infrastructure; would a consistent 20 hours/week meet your availability requirement, or is 20–30 hours a hard minimum every week?

6
  1. Project walkthroughs:
    https://www.tella.tv/playlist/cm5kf4s55000u09kv1ixl7v5a/public
    Portfolio: Selected work | LFSA Digital
  2. My standard implementation rate is $50/hour. The $300 receptionist installation is below my normal rate for a 9 to 11-hour implementation, but I would consider the model if the recurring $100 monthly share and support cap are contractually preserved as described. Otherwise, I would propose $450 per receptionist installation.
  3. I can initially commit 20 hours per week, with flexibility around planned installation windows.
7
  1. Things I have built and deployed

Live multi-tenant SaaS: https://tarkabot.online — WhatsApp ordering system
for restaurants, React/Next.js on Vercel, FastAPI backend, Supabase
Postgres with row-level security on every table and 48 migrations, 32
automated tests, CI on GitHub Actions.

Live site: https://mahsaanullah.online

Code: https://github.com/MAhsaanUllah — WARAQ-AI (multi-tenant RAG, 55
tests), agent_friday (LangGraph agent with human approval gates + MCP,
68 tests + CI), ScreenOS (FastAPI + LLM scoring pipeline with guardrails
and evaluation).

Automation: 15-node n8n workflow for restaurant order intake, a Gmail +
Gemini pipeline that drafts replies with human approval, and a voice agent
wired through n8n and VAPI.

  1. Rate expectations

Your fixed fees work for me as stated ($300 per install, $199 add-on) with
the recurring monthly share as the main value. Past the support window,
$95/hr is fine.

  1. Availability

20 hours a week to start, up to 30 during install weeks. Pakistan (UTC+5),
so your business day is my evening and night.

8

Hi All,

This req is filled.

9

Hi Matthew, I’m interested in the contract implementer role.

My background is mainly in automation, AI integrations, APIs and webhooks, and I also have experience building and deploying web applications, including an AI-powered site hosted with Vercel.

I’m comfortable working from a written implementation runbook, troubleshooting integrations and handling client-side setup and support.

I’ll send over my GitHub, portfolio and availability by email. Thanks!

11

Hi Matthew,
Your setup is a good match for the kind of implementation work I handle. Since the core engine and runbook are already in place, I can focus on getting each client environment deployed correctly, validating the integrations, and supporting what I install afterward.

1. Relevant work:
My experience includes Next.js, TypeScript, APIs/webhooks, PostgreSQL, GitHub deployments, cloud environments, AI integrations, and third-party services. I’m also comfortable working through an established provisioning process rather than changing a working architecture unnecessarily.

2. Rate expectations:
I’m open to starting with your $150–$200 paid screening task. For ongoing work, the fixed install + recurring support model can work for me, provided we confirm the expected scope and support boundaries during the trial.

3. Weekly availability:
[ 30hours/week ]

I’d be happy to start with the paid screening task and let the actual implementation work determine whether we’re a good long-term fit.

12

Hi Matthew,

1. Built and deployed:
— Multi-agent lead qualification system, Gulf real estate client: GitHub - ai-with-abdullah/prime-estate-chat · GitHub
— WhatsApp Business API booking agent, UK dental clinic, live client traffic, calendar integration, escalation to human
— Menu-grounded ordering agent, restaurant client in Oman.

Straight answer on your stack: I’ve shipped conversational agents to paying clients and I work to a runbook, but Next.js/TypeScript, Supabase RLS, Cloudflare Workers and Twilio Voice are not things I’ve deployed. My depth is Python, n8n, LangChain and the LLM layer. If you need someone already fluent in that stack from day one, I’m not him. If the screening task is partly a test of how fast someone picks up your conventions, I’d take it on.

2. Rate: your fixed fees work for me as published. I’d want to confirm the recurring share applies the same way to a non-US contractor.

3. Availability: 25 hrs/week, UTC+5, with 09:00–14:00 ET overlap.

One question before the screening task: I’m based in Pakistan, so I’d file a W-8BEN rather than a W-9. Does that work for you?

Muhammad Abdullah · abdullah.tech.ai@gmail.com

13

Hi Matthew,

Saw your post for the contract implementer role. I specialize in deploying production-ready AI intake and qualification architectures that bridge webhooks, LLM reasoning, and downstream tooling.

Recently deployed a complete pipeline using n8n + AI: • Ingests client intake via webhooks with zero latency • Routes payloads to an LLM node to evaluate intent, budget ($3k+ priority tagging), and qualification scores (90+) • Auto-syncs structured reasoning into Google Sheets CRM • Automatically dispatches dynamic Calendly invites within 5 seconds to lock in booking calls

Recorded a quick 45-second demo showing the live workflow: Client Project Intake G - Google Vids

Comfortable handling API integrations, webhook routing, and debugging production workflows. Available to start immediately and hop on a quick sync if you’d like to review the setup.

Best,

Niteesh Kumar

14

Hi Matthew,

I’m reaching out regarding your Contract Implementer opportunity for Bridge Builders AI.

I build AI automation and backend systems using Python, FastAPI, n8n, APIs, webhooks, SQL, and AI/LLM integrations. Two relevant projects you can review:

SupportPilot AI:

AI Lead Management Agent:

These projects demonstrate API/webhook integrations, structured data handling, validation, automated testing, error handling, and AI workflow orchestration.

I’m comfortable with your stated fixed-fee project structure and the $150–$250 paid screening task. I’m available around 10–20 hours per week remotely from India.

I don’t want to overstate experience with every tool in your stack, but I’m comfortable working from technical runbooks and documentation, learning any platform-specific setup required, and personally owning implementation, debugging, testing, and delivery.

Happy to complete the paid screening task if you think my background is a fit.

Best,
Ronak Patel
GitHub: ronakkhunt28-create (Ronak Patel) · GitHub
LinkedIn: https://www.linkedin.com/in/ronak-patel-72039b3ba

15

Hi Matthew,

I run a small automation studio (jnxautomation.pl) and a lot of my work is exactly this pattern: a working engine plus a written runbook, deployed client by client and then kept alive. Last month I shipped a lead-intake automation for a B2B service company (mailbox routing, dedup, error alerts) and I work daily with n8n/Make, Python, REST/webhook integrations. On your stack I work with AI-assisted implementation and verify the result against your own gates before delivery: build and render, consent gate end to end, Twilio failover, secrets server-only, live call test.

The way I would add value beyond executing the runbook: the first install I document step by step, and that document becomes the checklist for the next one, so install five costs a fraction of install one. I treat the verification gates as the definition of done, not as a formality, and I keep the client walkthrough in the same place as the install notes.

Two questions before we start. What pace and volume of installs do you expect per month, and what is the rate per install plus the monthly support cap? Could you share one reference client repo and the runbook, so I can dry-run the flow and show you how I work before the first paid installation?

I’m in Warsaw (CET, overlapping your mornings) and I work asynchronously with short written updates. I can take a paid pilot install this week.

Best regards,
Jan Noga
jnxautomation.pl

16

Hi,

I build and ship on most of that stack already, so I will be specific about what I have run in production and what I have not.

Next.js, Supabase and Vercel are my daily tools. I built WIVO, an on-demand delivery platform for Yaoundé and Douala, and published it on the App Store and Google Play. It handles customers, couriers and merchants, and it takes Mobile Money payments, which fail, duplicate or arrive twenty minutes late often enough that the whole order lifecycle had to be designed around unreliable confirmations. I also lead a production intranet used daily by staff, teachers and students, with separate development, test and production environments and automated backups.

On the messaging and AI side, I built a WhatsApp assistant on Meta’s Cloud API that answers customers through a cascade of language models with a fallback when one fails, keeps conversation state in Postgres, and reads screenshots people send as proof of payment. That is the same shape of work as an AI phone line: webhooks in, state kept somewhere reliable, a model in the middle, and something sensible to say when a provider times out.

What I have not shipped yet: Twilio specifically, ElevenLabs, and Cloudflare Workers. I have worked against the equivalent messaging APIs and I deploy on Vercel rather than Workers. I would rather say that now than discover it on your first client install.

Repeatable installs and a capped monthly support commitment suit how I work. I am in Tunis (UTC+1), I can overlap with US mornings, and I can be paid through Deel or by bank transfer.

Happy to walk through the delivery pipeline of any of the above if useful.

Rick Kendem — EOWEB

17

1. Things I’ve built and deployed
• Live client website/product ecosystem: https://mediumturquoise-weasel-752855.hostingersite.com/
• Built and supported a live GoHighLevel-based Voice AI lead protection system covering inbound call handling, AI qualification, CRM capture, booking logic, SMS notifications, multilingual voice agents, testing, and handover documentation.
• Built real-estate CRM automations covering Meta lead capture, lead routing, appointment booking, reminders, custom fields, permissions, WhatsApp/SMS workflows, webhooks, and integrations.
• Completed A2P 10DLC implementation and compliance work including brand/campaign setup, opt-in flow review, Privacy Policy/Terms alignment, message samples, and registration troubleshooting.
• Hands-on experience with Twilio, Voice AI, APIs/webhooks, Make, n8n, GoHighLevel, WhatsApp, SMS, and client-facing automation systems. Additional private workflow examples can be shared where client confidentiality allows.

2. Rate expectations
I’m comfortable starting with the paid $150–$250 screening task and working within the fixed install fees you listed. I understand the economics are weighted toward the recurring monthly share and long-term book of live clients, and I’m comfortable evaluating that structure through the first implementation.

3. Weekly availability
20–30 hours per week.

18

Interested — and this maps onto my daily stack: Next.js/TypeScript on Vercel with Supabase (Postgres, Auth, RLS, migrations), plus GitHub Actions.

Straight answer on the vendor side: Next.js + Vercel + Supabase I do end to end. Twilio Voice/TwiML, ElevenLabs Agents, Resend and the Cloudflare Worker media gate I have not shipped in production yet — but they are config work against a written runbook, and a runbook is exactly how I like to work. I’d rather tell you that now than discover it on a client call.

Proof of work: https://landing-showcase-gamma.vercel.app/ and mohsenjfar (Mohsen) · GitHub

The pay structure works for me — I’d rather earn toward the recurring share than a one-off fee.

One logistics note so neither of us wastes time: I’m based in Iran, so I invoice internationally in USDC. If that’s a blocker for your accounting, say so now and no hard feelings. If it isn’t, I can start on the runbook the same day you send it.

19

Hi Matthew,

The three things you asked for:

  1. BUILT AND DEPLOYED

• +506 6060 0870 on WhatsApp — my company’s AI sales agent, live on the WhatsApp
Cloud API. Write to it in English; it answers in whatever language you use.
It’s a sales agent, so it will ask about your business — tell it you’re
evaluating me and it steps aside cleanly. Claude for the reasoning, Google
Calendar for the booking, Google Sheets as the CRM, email alert to my team on
every lead. It’s the closest thing I have to your AI Customer Assistant, and
it runs on a real phone number rather than a simulated chat window.

• https://veterinaria-nuva.vercel.app — the same engine reskinned for a
veterinary clinic. Books into Google Calendar, sells from a 10-product catalog
(sends the product photo in-chat), logs every interaction to Sheets.

• https://barberia-nuva.vercel.app — same engine again, with real double-booking
protection. Reskinning that engine per client is most of what I do.

• https://camilo-woodcock.vercel.app — my portfolio, Next.js on Vercel.

• https://nuva-tour-jaco.vercel.app — 360° property tour running on a
hand-written equirectangular WebGL shader, no library. Included because it
shows I can work at the level below the framework when an install goes wrong.

Everything above is deployed on Vercel through the CLI — that part of your
pipeline is my daily workflow.

Where I’m strong: Next.js, TypeScript, Vercel CLI, npm workspaces, deployment
and verification, and the AI-agent side of the stack.

Where I’d be learning on your runbook: Twilio Voice/TwiML, ElevenLabs Agents and
Cloudflare Workers/Wrangler. I’ve read the docs, I haven’t shipped them in
production. I’d rather you know that now than find out during the screening
task. Supabase I’ve specced (Postgres, Auth, RLS, migrations) but not shipped.

  1. RATE EXPECTATIONS

Your fixed fees work for me as posted — $300 per install, $199 for the assistant
add-on, monthly recurring as listed. $95/hr for support beyond the cap is fair
and I’m not looking to renegotiate it.

I’d like to take the paid screening task. Quote whatever’s standard for you in
the $150–250 range; I’m more interested in showing you the work than in the
number.

  1. WEEKLY AVAILABILITY

20–25 hours a week, sustained and predictable rather than bursty.

I’m in Costa Rica (UTC−6), so I overlap Eastern US business hours completely —
same-day turnaround during your working hours, not overnight.

One note on how I work: I’m strongest async and in writing. Written handovers
plus recorded walkthroughs are my default, and I saw that works for you. My
English is solid in writing; I’d rather hand you a clear runbook than a phone
call.

— Camilo Woodcock
camilow14@gmail.com

20

Camilo,
Straight answer on the implementer role: no, for now. The installs run on Twilio Voice and ElevenLabs Agents, and you have shipped neither.
There may be something else. It depends on two answers below.

  1. Have you ever put a voice layer in front of that engine? Any telephony, not only Twilio. If you have, tell me what broke.
  2. The barbershop build says real double-booking protection. How does it actually work? I want the mechanism, not the feature.
    I am asking because a booking agent on a live number, with a CRM log and an alert on every lead, is the nearest thing anyone has sent me to work I have coming, building this functionality now myself while running interviews for tech sales appointment/lead gen and a ton of other things competing for my attention.
    The voice layer is the gap, and those two answers tell me whether it is a gap or a wall.
    Matthew Ferrill
    Bridge Builders AI LLC
21

1 — Built and deployed

Everything I have shipped is here, and each card opens the live site: https://villalpando.dev

The ones closest to your runbook:

  • Relojes Exclusivos BA — Next.js 16 + TypeScript on Vercel, Supabase with Postgres, Auth, RLS and 53 versioned migrations. Live bidding over websockets, a purchase advisor on the Claude API, i18n in three languages. A second brand runs on the same database behind a per-site guardrail, so one engine serves two separate client-facing storefronts.
  • Granja Lo del Gaita — Next.js 16 + Supabase (RLS, Storage, Edge Functions), MercadoPago checkout with webhook, Resend transactional email, delivery-zone geofencing, CI on GitHub Actions running lint, typecheck and tests before every deploy.
  • ASISCOM — WordPress to Next.js migration with no email downtime, custom signed-session admin, Google Places API, GA4/GTM and technical SEO. At the end I handed the Vercel and Supabase projects over to the client’s own account.
  • Red Contenido IA (private) — 13-package pnpm monorepo in TypeScript and Python, 22 n8n workflows in production, seven Claude API agents, ElevenLabs voiceover, Docker on Railway, GitHub Actions CI and Sentry.

Per-client provisioning, applying migrations, keeping an env contract straight and deploying prebuilt output through the Vercel CLI to projects that are not connected to Git are things I already do weekly.

Where I am light, said up front: I have not shipped Cloudflare Workers or Wrangler — I use Cloudflare for DNS only. Twilio Voice I have wired into a voice-agent prototype, not into production. ElevenLabs, Supabase, Resend, Vercel CLI and GitHub Actions I use in production today.

2 — Rate

Your fixed fees work as posted: $300 / $300 / $300 / $199, and $95/hr past the support cap. I read your framing on the install fee and I agree with it — the recurring share per live client is what makes this worth doing. I am interested in building a book of live clients, not in a string of gigs.

Happy to take the paid screening task at $150.

3 — Availability

20 to 25 hours a week sustained, and up to 30 during a heavy install week. I am in Argentina (UTC-3), which overlaps the entire US business day.

I work to a written runbook and document what I hand over. If it helps, I can walk one of the sites above end to end — repo layout, env contract, migrations, deploy — so you can judge the conventions rather than the screenshots.

— Agustín Villalpando