PamiTC Job Board - Posting Details

Title: Computer vision / machine learning scientistPosted: May 27, 2015
Company/Institution: Moodstocks
Location: Paris
Department: R&D

Description: At Moodstocks we are building the Vision platform for the IoT, in order to make connected devices smarter by giving them eyes. +8,800 developers worldwide signed up so far. +40 tier 1 retailers, publishers and brands (like L'Occitane, Gucci, Vuitton, Rayban, etc) rely on our API / SDKs to power visual search within their mobile apps.

We are a team of passionate engineers and PhD-s working hard to turn deep learning algorithms into *real-world* mobile applications:

- our main focus is on visual search (a.k.a. search by example) targeting any kind of real world objects from fashion to furniture which translates into fine grained visual recognition,
- our specificity is the on-device first approach: our pipeline is made to work right on the device to avoid network latencies at query time and thus ensure instant recognition.

In terms of tools / languages:

- we work with Torch for prototyping and thus use Lua / C but also Python when appropriate
- we leverage GPU computations via Torch CUDA extensions for our intensive learning stages
- we collaborate on GitHub / Flowdock
- we deploy on Amazon Web Services (Linux) for production
- we use open source software and do our best to contribute back to the community

In this context we are looking for skilled engineers / scientists with a deep interest in computer vision / {machine, deep} learning to help us improve our pipeline:

- review, comment, select and prototype ideas from public research papers
- suggest ideas combining different approaches
- improve our internal frameworks
- extend them to work with new verticals
- increase the accuracy of our networks
- enhance the speed and reduce the footprint

Application Instructions: