Issa Loubani

Issa Loubani

AI Engineer — Voice & Real-Time Speech

Download PDFBack to site

Profile

AI engineer who builds voice agents that answer real phone calls. Started the AI practice at a MENA telecom software vendor single-handedly and now owns the voice platform it turned into. Unusual combination of real-time audio, carrier telephony, and native Arabic speech evaluation. Five years shipping software, four of them at the same company across three roles. BE in Computer and Communication Engineering, GPA 3.8/4, with a machine-learning capstone in 2021.

Experience

AI Engineer, Voice & Real-Time Speech · Apliman Technologies

Feb 2025 — present

Started the AI work at Apliman alone while in R&D. Customer demand outgrew one person and the company formed an AI team around it. I own the voice agent platform and onboard the engineers joining the team.

  • Built the AI voice agent platform sold to telecom and enterprise customers across MENA. Sole initial developer, current maintainer.
  • Shipped V1: agent management UI, WebRTC and AudioSocket transports, LLM tool and function calling, knowledge base management.
  • Put agents on live phone calls over carrier-grade SIP, with Asterisk, AudioSocket and DTMF for IVR-style flows. Built the transport layer beneath: Coturn, RTPEngine, ICE, NAT traversal, SIPp load testing.
  • Delivered assistants for banking, mobile money and hospital appointment booking, in bilingual Arabic and English and in French-first flows.
  • Deep Pipecat work in Python and FastAPI: frame architecture, custom processors and transports. Benchmarked self-hosted speech stacks (Faster Whisper, XTTS v2) against hosted APIs on latency, cost and Arabic accuracy, running the Arabic evaluation personally as a native speaker.
  • Onboard and mentor new AI engineers and interns. Building V2: multi-tenant architecture, analytics, encryption and security.

R&D Engineer · Apliman Technologies

Jan 2024 — Feb 2025

Promoted into R&D to evaluate emerging technology and turn it into tooling the engineering teams could use. The AI work started here as a side track and became the job.

  • Prototyped an LLM tool that answered questions about the codebase and generated code in context.
  • Evaluated Kubernetes, gRPC, and NLP and LLM tooling for internal use.
  • Supported product teams in Flutter through delivery crunches.

Full Stack Developer · Apliman Technologies

Sep 2022 — Jan 2024
  • Built web and mobile applications in Flutter and REST APIs in Spring Boot.
  • Deployed with Docker and Tomcat. Wrote a reusable Flutter library released for other teams.

Flutter Developer · Easysoft ME

Sep 2021 — Aug 2022
  • Built and maintained Android and iOS applications, published to Google Play and the App Store.
  • Wrote reusable Flutter libraries and built data visualisation features from Adobe XD designs.

Published work

Skills

Voice & real-time
Pipecat, WebRTC, SIP, Asterisk, AudioSocket, DTMF, Coturn, RTPEngine, ICE, SIPp
AI & speech
STT, TTS, ASR, LLM tool calling, Faster Whisper, XTTS v2, LoRA fine-tuning, YOLO
Backend
Python, FastAPI, Spring Boot, Docker, REST, gRPC
Frontend & mobile
Flutter (web, iOS, Android), Dart, Kotlin, Java
Languages
Arabic (native, Levantine & MSA), English (fluent), French

Education

BE, Computer and Communication Engineering · Lebanese International University

2017 — 2021

GPA 3.8/4. Dean's and President's lists. Capstone applied machine learning to image compression.

Certifications