Besedo Tech Stack

AI-powered content moderation with human review for marketplaces and social platforms

Software Development Stockholm, Stockholm County 201–500 employees Founded 2002 Privately Held

Besedo operates a hybrid content moderation platform combining PyTorch models, real-time inference (FastAPI, SageMaker), and human review workflows. The tech stack and active projects reveal a company scaling ML production infrastructure—training pipelines, model drift monitoring, and high-performance inference services dominate their roadmap. Hiring is engineering-heavy (9 of 15 roles) and weighted toward senior talent, suggesting they're building infrastructure and automation rather than expanding review capacity.

Tech Stack 19 technologies

Core StackPython PyTorch scikit-learn GitLab CI/CD Docker Java Spring Node.js JavaScript AWS Kubernetes Elasticsearch TypeScript FastAPI SageMaker ONNX TensorRT Spring Boot OpenSearch

AdoptingKubernetes AWS FastAPI SageMaker

What Besedo Is Building

◆Challenges

Ensuring reliability and scalability of ml production
Fraud detection
Optimizing model performance and efficiency
Moderating large volumes of content
Model drift monitoring
Blocking fraud attempts
Policy compliance
Customer dispute resolution
Scams and fraud
Ml productionization

▲Active Projects

Ml production lifecycle management
Training and retraining pipelines
High-performance inference services
Production lifecycle of ml systems
Improving publication rules
Fraud monitoring
Training pipeline development
Inference service development
Microservices development for implio platform
Infrastructure ci/cd set-up

Hiring Activity

Accelerating15 roles · 7 in 30d

Department

Engineering

Support

Sales

Seniority

Senior

Junior

Company intelligence

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

Get started free

About Besedo

Besedo provides content moderation services for marketplaces, social platforms, and classifieds sites. The offering pairs machine learning models (PyTorch, scikit-learn) with human moderators to flag and resolve policy violations, fraud, and scams at scale. They operate across Sweden, Malta, and Romania, with a team of 201–500 spanning engineering, support, and sales. The platform handles real-time classification and policy enforcement, with ongoing work on model retraining, inference optimization, and fraud detection accuracy.

HeadquartersStockholm, Stockholm County

Company Size201–500 employees

Founded2002

Hiring MarketsMalta, Romania, Sweden

Frequently Asked Questions

What machine learning framework does Besedo use?

PyTorch for model training, ONNX and TensorRT for inference optimization, and scikit-learn for classical ML. They deploy via AWS SageMaker and FastAPI for real-time serving.

What is Besedo's main engineering focus?

ML production infrastructure: training and retraining pipelines, high-performance inference services, model drift monitoring, and microservices development for their moderation platform.