MLOps London November - Talks on Architectural Design and Computer Vision

📽️ Livestream: https://www.youtube.com/watch?v=2h-IpI-BXqg
🧑 In Person: https://forms.gle/syYGRS6cAz5GQtzk9
***

MLOps London is back again in November with more talks on production machine learning, DevOps and Data Science. The plan, as usual, is to run another hybrid event so please come along in person if you're local. If you're further afield you will still be able to join the stream.

***

AGENDA:
⏱️ 6.00pm onwards
🍺 Arrival, drinks and networking

⏱️ 6.30pm
🎤 Kick off and welcome

⏱️ 6.40pm
🎤 Zero Shot Synthetic Image Recognition, from 3D asset to Embedding
🧑 Patric Fulop - Co-Founder & CTO @ Neurolabs

This talk will cover challenges & solutions for detecting physical products with zero shot image recognition across different use cases and with an ever changing catalogue.
Within the Consumer Product Goods (CPG) space, accurate object recognition is a challenging task due to the highly diverse real environments a long tail distribution in the real data available and a constantly changing product catalogue.

We present a synthetic image recognition pipeline that aims to solve the above challenges and present three main ingredients:
- a data generation engine that creates synthetic data from 3D models
- a domain agnostic vision module that creates synthetic data embeddings
- the MLOps tools used to deploy the vision module for clients to onboard new products & detect them in the real world.

⏱️ 7.25pm
🎤 Turbocharging architectural stakeholder engagement with ML
🧑 Michael Calver - CEO @ Amutri

Amutri is building a first of it's kind architectural design communication platform that allows architects to engage with the full range of stakeholders and local communities. This talk will present the wide range of machine learning use-cases involved in this product that enables stakeholders to work together to deliver higher quality and cost efficient architectural outcomes.

***
If you are attending in person please complete the registration form (link at top of this description)