Learn more about the magic behind BingeWiz

The Tech Stack Breakdown

Data preprocessing pipeline
Model training Model Serving
Experiment tracking, data versioning, model registry
Feature StorE
Model storage
Vector Similarity Search
Monitoring
Embedding/llm API

INNOVATION

What sets us apart

User-centric recommendations

We base our recommendations solely on user watch history, allowing us to train on unrelated datasets.

ENHANCED WITH LLM Embeddings

We use OpenAI’s API to learn embeddings from movie plot and other textual features, reducing overfitting, enhancing model results and enabling the model to recommend new or lesser-known movies.

IMPROVED Embedding Techniques

By concatenating ratings into user watch history before applying an attention layer, we supercharge our model metrics.

Best Practices in ML Operations

Key Steps for Effective Model Deployment and Performance Optimization
1

Clear requirements

We set budget-aligned limits on deployment from the start, avoiding unnecessary investments in design and models, quickly abandoning frameworks and models that would waste time.
2

Experiment tracking

Using ClearML, we document training runs with code changes, dataset versions, and hyperparams for easy result comparison and reproduction.
3

Model monitoring

Our system is continuously monitored with Prometheus and Grafana, ensuring production success. We define metrics to compare the live model to the offline evaluations, and measure speed, throughput and other metrics.

Read from the mind of our experts

Machine Learning
By Martin Barreto
How To get the most out of a recommendation engine
Read blog post
Machine Learning
By Mathias Claassen
TF recommenders: encode user watch history using openAI embeddings
Read blog post
Machine Learning
By Mathias Claassen
Deploying a Recommendation system using FastAPI, TF Serving and Feast
Read blog post
Machine Learning
By Maximo Lionello
ML Model monitoring: boosting performance & reliability
Read blog post