V

Verbex.ai Tech Stack

Voice AI agents powered by proprietary LLM stack and speech models

IT Services and IT Consulting Singapore 11–50 employees Founded 2017 Privately Held

Verbex builds voice-first AI agents for customer engagement, sales, and support using a custom LLM stack (Llama, Hugging Face Transformers, LangChain) paired with proprietary speech technology. The engineering-focused org is actively integrating advanced speech foundation models (SpeechGPT, Moshi, Llama-Omni) and scaling inference with TensorRT + vLLM, signaling rapid iteration on model quality and latency performance. Active projects span TTS/STT development, Kubernetes-based microservice deployment, and telephony integration—suggesting a maturing platform transitioning from research to production at scale.

Tech Stack 50 technologies

Core StackDocker Kubernetes Linux Python AWS LangChain RAG PostgreSQL MongoDB Kafka RabbitMQ Redis ArgoCD GitHub Actions Jenkins Bitbucket Terraform Ansible Prometheus Llama ONNX TensorRT vLLM Git Hugging Face Transformers LlamaIndex MCP GCP Elasticsearch, Logstash, Kibana Bash+20 more

What Verbex.ai Is Building

◆Challenges

Improving container-based microservice infrastructure
Continuous delivery & integration adoption
Backup and recovery design

▲Active Projects

Develop cutting-edge tts, stt, llm solutions and conversational ai systems
Integrate advanced speech foundation models such as speechgpt, moshi, llama-omni
Deploy research models at scale using docker and kubernetes
Container-based microservice infrastructure improvement
Continuous delivery & integration deployment
Pbx system installation and configuration
Test automation framework for web and backend services
Ci/cd pipeline integration for automated testing
Telephony integration testing

Hiring Activity

Minimal5 roles · 0 in 30d

Department

Engineering

4

Seniority

Mid

2

Senior

2

Company intelligence

Find more companies like Verbex.ai by tech stack, pain points and active projects

Get started free

About Verbex.ai

Verbex develops conversational AI agents that interact via voice across phone, web, mobile, and IoT channels. The platform targets enterprise use cases—SDR automation, support agents, appointment booking, and conversational payments—across finance, healthcare, and other regulated verticals. Built on a proprietary LLM foundation and speech synthesis/recognition stack, agents handle high throughput (2M+ calls/hour) with low latency and operate across multiple jurisdictions (Japan, Bangladesh). Founded in 2017 and based in Singapore, the company holds 24 patents and operates a compact, engineering-led team.

HeadquartersSingapore

Company Size11–50 employees

Founded2017

Hiring MarketsBangladesh

Frequently Asked Questions

What tech stack does Verbex use for voice AI?

Llama, Hugging Face Transformers, LangChain, ONNX, TensorRT, and vLLM for inference; Docker and Kubernetes for deployment; PostgreSQL and MongoDB for persistence; Kafka and RabbitMQ for messaging.

What are Verbex's main projects?

Core work includes TTS/STT and LLM development, integrating speech foundation models, Kubernetes microservice infrastructure, CI/CD pipeline automation, and telephony system integration testing.