MUHAMMAD AHMAD SIDDIQUI

Full Stack & AI Engineer

000

Available for work·Book a call
All Projects/Sentire Systems

Full Stack Developer

Sentire Systems

AI voice platform that gives service businesses a 24/7 receptionist, answers calls, books appointments, and recovers missed callers automatically

Next.jsFastAPIOpenAI GPT-4oLangChainDeepgramElevenLabsTwilioSupabaseGoogle Calendar APIRedisAWS
Sentire Systems

Overview

Sentire Systems is a production-grade multi-tenant AI voice platform designed to automate customer communication, inbound calls, and appointment scheduling for service-based businesses. The platform provides businesses with dedicated AI voice agents that answer calls 24/7, understand customer intent through natural conversation, schedule appointments in real time, and recover missed callers via SMS, all without a human receptionist. Traditional businesses lose potential customers due to missed calls, delayed responses, and manual scheduling. Sentire replaces those workflows entirely with an always-on, intelligent voice assistant.

The Challenge

Delivering natural voice conversations required sub-2-second end-to-end latency across a full real-time pipeline: telephony routing, speech-to-text, LLM reasoning, text-to-speech, and audio playback. Supporting multiple businesses on one platform demanded strict tenant-level data isolation with zero cross-tenant leakage. The system also needed to handle 30+ concurrent calls per instance without degrading response quality, and support cross-industry flexibility, salons, clinics, professional services, through a configurable agent architecture rather than separate codebases.

How We Solved It

1

Voice Interaction Layer

Real-time voice pipeline: Twilio manages business phone numbers and call routing, audio streams over WebSockets, Deepgram converts customer speech to text, OpenAI GPT-4o processes conversation, and ElevenLabs generates natural voice responses. Optimized at every stage to achieve sub-2-second round-trip conversational latency.

2

AI Conversation Engine

Powered by OpenAI GPT-4o with LangChain orchestration. The engine receives live call audio, converts speech to text, detects customer intent, maintains multi-turn conversation context, executes business workflows (scheduling, FAQs, escalation), and generates responses, all within a single streaming pipeline to minimize delay.

3

Appointment Automation

Deep Google Calendar integration enables real-time availability checks, conflict-free booking, business-hours enforcement, and calendar synchronization. Customers complete full appointment bookings entirely through voice, zero clicks, zero forms, zero wait.

4

Multi-Tenant SaaS Architecture

Each business receives a dedicated phone number, isolated AI configuration, separate customer data, independent dashboard, and custom branding. Supabase Row-Level Security policies enforce complete tenant separation at the database layer, ensuring no cross-tenant data access is possible.

5

Dashboard & Management System

A full business dashboard covers call history, conversation transcripts, audio recordings, appointment analytics, AI personality configuration, business hour settings, agent prompt management, SMS logs, missed call recovery, and branding customization, giving owners complete visibility without sitting at the front desk.

6

Missed Call Recovery & SMS Automation

When a caller hangs up before connecting, the platform automatically sends an AI-powered SMS with available times, handles the customer's reply, and confirms the booking, converting missed calls into appointments entirely over text with no human intervention.

7

Operations Integrations

Per-tenant connectors plug the agent into the tools a business already runs: Toast, Square, and Clover POS for live menu pricing and order sync, Star CloudPRNT for printing phone orders straight to the kitchen, and Stripe for secure payment links sent by text. A central Sentire Engine standardizes webhook payloads from each integration and streams updates to calendars and phone platforms in real time.

Platform & Landing

Business Dashboard

Services & Operations

Results & Impact

Sub-2s Conversational Latency

End-to-end voice pipeline optimized across telephony, STT, LLM reasoning, and TTS to deliver human-like conversation speed.

30+ Concurrent Calls

Horizontal scaling architecture handles 30+ simultaneous voice calls per instance without degrading response quality or latency.

24/7 Call Coverage

Dedicated AI voice agent answers every inbound call around the clock, no missed calls, no voicemail, no wait times.

Zero Double Bookings

Real-time Google Calendar availability checks prevent scheduling conflicts, every slot verified before a booking is confirmed.

Missed Call SMS Recovery

Automatically follows up on missed calls with available times over SMS and converts replies into confirmed appointments.

Full Multi-Tenant Isolation

Supabase Row-Level Security ensures complete per-tenant data separation across all businesses on the shared platform.

Original Project

Client's product site

Company website

sentiresystems.com (opens in a new tab)

Company website for the AI voice platform built here. Some features described in this case study may not yet appear on the live site.

Technical Architecture

FrontendNext.js + Tailwind CSS, landing site, client login, business dashboard (call logs, appointments, analytics, agent config, SMS logs, branding)
BackendPython FastAPI, call orchestration, WebSocket audio streaming, multi-tenant management, appointment workflows, SMS automation
AI / LLMOpenAI GPT-4o + LangChain, intent detection, multi-turn conversation, business rule execution, response generation
Voice PipelineDeepgram, real-time speech-to-text; ElevenLabs, natural text-to-speech voice generation; Twilio, telephony, phone number management, WebSocket audio routing
SchedulingGoogle Calendar API, real-time availability checks, conflict-free booking, business-hours enforcement, calendar sync
Database & AuthSupabase PostgreSQL + Row-Level Security, multi-tenant data isolation, call records, transcripts, appointments; Supabase Auth
Caching & InfraRedis, session caching, real-time state management; AWS, scalable cloud deployment with horizontal scaling for concurrent call handling

Have a similar project?

I'm open to freelance projects and full-time roles. Let's build something great together.