UNITH Tech Stack

AI-powered digital humans for conversational customer engagement

Software Development Barcelona, Catalonia 11–50 employees Founded 2014 Public Company

UNITH builds conversational AI agents that synthesize video in real time, pairing PyTorch + TensorFlow inference with AWS Bedrock and OpenAI integrations. The tech stack (gRPC, Redis, RabbitMQ, FastAPI) and active project list (video synthesis pipeline, multi-document RAG, model versioning) reveal a company solving for low-latency, stateful conversations at scale — a harder problem than text-only chatbots. Engineering-focused hiring (mostly senior roles) suggests they're solving deep inference and backend scaling challenges rather than expanding sales.

Tech Stack 24 technologies

Core StackPython PyTorch TensorFlow AWS Docker Kubernetes gRPC Redis RabbitMQ FastAPI Flask Go OpenAI AWS Lambda Zapier n8n React JavaScript TorchServe ONNX Runtime AWS Bedrock REST HTML CSS

What UNITH Is Building

◆Challenges

Reducing inference latency for real-time video synthesis
Scaling backend systems for conversational ai
Real-time streaming architecture
Onboarding process
Demo setup
Internal tool automation

▲Active Projects

Real-time video synthesis pipeline
Avatar customization capabilities
Backend systems for conversational ai
Multi-document knowledge bases using aws bedrock
Llm api integration and rag pipelines
Model versioning and a/b testing
Video synthesis api design
Internal tool automation
Integration development
Usage tracking and data collection

Hiring Activity

Accelerating5 roles · 3 in 30d

Department

Engineering

Seniority

Senior

Intern

Company intelligence

Find more companies like UNITH by tech stack, pain points and active projects

Get started free

About UNITH

UNITH is a public company (ASX: UNT) founded in 2014 and headquartered in Barcelona. The company develops digital humans — conversational AI agents capable of generating video responses in real time — targeted at customer engagement, sales, and support use cases. The platform combines LLM reasoning (via OpenAI and AWS Bedrock) with video synthesis, requiring orchestration across inference engines, streaming backends, and API layers. Current hiring is concentrated in Spain and focused on senior engineering talent.

HeadquartersBarcelona, Catalonia

Company Size11–50 employees

Founded2014

Hiring MarketsSpain

Frequently Asked Questions

What AI models and frameworks does UNITH use?

PyTorch, TensorFlow, AWS Bedrock, and OpenAI APIs. They also use ONNX Runtime and TorchServe for model serving, and n8n/Zapier for workflow orchestration.

What is UNITH working on?

Real-time video synthesis pipelines, avatar customization, LLM-powered conversational backends, multi-document knowledge bases via AWS Bedrock, and RAG-based API integrations. Current focus areas include reducing inference latency and scaling backend systems.