Issa Loubani

Voice agents that
answer real phone calls

AI engineer, Beirut. I build real-time systems: software that perceives something as it happens and answers before the moment has passed. Right now that means voice agents on carrier telephony, in Arabic and English.

PipecatWebRTCSIPAsteriskAudioSocketSTTTTSLLMFastAPIPythonArabic speechDTMFCoturnRTPEngineSIPpICEPipecatWebRTCSIPAsteriskAudioSocketSTTTTSLLMFastAPIPythonArabic speechDTMFCoturnRTPEngineSIPpICE
/ Published — live on public platforms

Shipped, listed, and downloaded by strangers

JetBrains Marketplace

glb-viewer

An IntelliJ plugin that opens and inspects 3D GLB models inside the IDE, so you stop alt-tabbing to a separate viewer.

Downloads
7,931
Rating
3.2 / 5
Plugin ID
22024
The red hand riding the progress bar inside the IDE
JetBrains Marketplace

omori-progress-bar

Replaces the IntelliJ progress bar with a themed one. Not useful. Published anyway, and people install it.

Plugin ID
33141
Scope
All JetBrains IDEs
Live — arabizi_transliterator v1.0.0

This is the actual published package, compiled from Dart to JavaScript and running in your browser. Type Arabizi below.

…
idle — package not loaded yet
/ Work — production and personal

Systems that have to answer in time

AplimanSince 2025

Voice agent platform

AI agents that answer real phone calls, over carrier telephony rather than a browser tab. Bilingual Arabic and English, in production across MENA.

RoleSole initial developer, current maintainer
StatusIn production
StackPipecat, Python, FastAPI, WebRTC, SIP, Asterisk, STT/TTS
Turn budget — target allocation800 ms
Endpointing / VAD120ms
Transcription180ms
Model300ms
Synthesis150ms
Network / SIP50ms
00:00.000OKcaller speech ends, endpoint detected
00:00.300RUNtranscript ready, dispatching to model
00:00.600RUNfirst token, synthesis starts streaming
00:00.800OKfirst audio frame reaches the caller
00:00.801HOLDanything past here and the pause is audible
Computer vision2024

Real-time gameplay analysis

Fine-tuned segmentation models that read live gameplay and call out what to do next: traversable geometry in one game, enemy positions and cover in another.

RolePersonal project
StackYOLO, Segmentation, Real-time, PyTorch
Capturepending
footage not uploaded
LLM agent2025

Live coaching agent for a 1990 NES game

An assistant that reads a running NES game's memory and gives tactical advice mid-match, based on score, stamina, and who is on the pitch.

RolePersonal project
StackLLM, Groq, Memory reading, Real-time
Capturepending
footage not uploaded
/ Also — archive
/ About
readme

I started the AI work at my company on my own. Customers liked what came back, demand outgrew one person, and the company built an AI team around it. I wrote the voice agent platform, I still maintain it, and I onboard the engineers joining the team.

Most of the work is Pipecat and Python. The part fewer engineers have done is the other end of the line: carrier SIP, AudioSocket, Asterisk, DTMF, WebRTC through NAT. A voice agent is a demo until it survives a real telephony network.

I am a native Arabic speaker, so I do the Arabic evaluation myself. Whether a synthesised voice reads a phone number correctly, whether it drifts out of dialect, whether a transcript is right or merely plausible. Benchmark scores catch none of that.

Before voice: three years of Flutter across mobile and web, and a stretch on WebRTC and TURN infrastructure. That background is why the telephony side does not scare me.

specv2026
RoleAI Engineer, voice & real-time speech
Since2021 — 5 years
CorePython, FastAPI, Pipecat
TransportWebRTC, SIP, Asterisk, AudioSocket
SpeechSTT, TTS, ASR, LLM tool calling
BeforeFlutter, Kotlin, Spring Boot
ArabicNative — Levantine & MSA
EnglishFluent
FrenchConversational
BasedBeirut, Lebanon
Open toRemote & Gulf
/ Contact
issa.loubani.5@gmail.com