We’re a team of 4 people people working on a natural language recommendation engine, powered by Bert. The hackathon team limit is 6, so we were wondering if there’s any people who may be a great fit for the team. Here are few areas of the project (experience in this is definitely not required).

-Network/Graph data processing/cleaning/preparation.

-Negative sampling data pipelines with Tf.data.

-Incorporating negative sample/candidate sampling into Keras.

-FAISS for embedding similarity lookup, millions of embeddings

-HuggingFace’s Tensorflow2.0/Keras models.

-TPU training with Keras.

Here’s a link to the hackathon

https://tfworld.devpost.com/

If you’re interested, pm me so that we can share backgrounds.