Lyst is a technology platform that revolutionises the way people shop for fashion. We connect millions of consumers globally with the world’s leading fashion designers and stores, giving them a simpler, more engaging and more effective shopping experience. Lyst has grown over 300% every year since launch in 2011 and has raised over $60M from top-tier investors including Accel, DFJ, Balderton and the teams behind LVMH, Michael Kors and Oscar de la Renta.

Lyst are looking for a Data Scientist to solve problems for our Atlas team in terms of Data Quality and De-Duplication that forms the beating heart of our product.

Given the amount of data ingested daily and the many different ways we process it, there is a large margin for error, so we're looking for a problem solver to build models for automation of these tasks at scale.

You'll be helping us solve myriad problems across our site: 

• Anomaly detection.
  • Specific focus on data health. You'll be analysing incoming data from spiders and augmenting or annotating the data to provide more information on possible errors and apply past learnings. 
• Analysing our existing catalog for errors and identify what these issues are and the scale of each one.
  • Develop more automated ways of improving the quality of our existing catalog. 
  • Use results from user research, stats from customer care and weekly QA metrics to direct focus and track / measure effect if required. 
• Confidence models for crowdsourcing.
  • Building out systems to learn from previously rejected / accepted candidate suggestions to (hopefully) more accurately determine if something is a potential duplicate or not.
    
• Further work on tag predictions for new products.
  • Working on colour pickers, looking into other meta data such as category and subcategory.

You'll be a voracious learner, keen to take ownership of projects and thrive on the autonomy that comes with them. You'll have the opportunity to write and build projects that will be at the heart of Lyst's platform and get to see them being used in the wild very rapidly.

You'll be part of a welcoming team of 5, who work closely together to improve the experience across our site through new research and close collaboration with other engineering teams across the business. 

We're a company built upon the principles of using data to make decisions and we need someone who shares our love for a data-driven approach to help us become the leading destination for fashion online.

You'll ideally be working somewhere solving similar problems, particularly in the fields of classification and data quality. We'll award big bonus points if you can bring experience creating convolutional neural nets. However our Data Science team have joined from Banks, research projects and more, so we're pretty relaxed!

You'll likely have a PhD in Computer Science or Electronics, but we also have people who trained in Economics, Maths, Physics and more. 

You'll be interested in the work that we are doing and be able to contribute to the further development via your experience with machine learning, NLP, stats modelling, deep learning and data mining in your previous research or commercial role. 

We're always looking for people who seek to demonstrate great Ownership, Humility, Talent, Impact and Drive. These behaviours help people to thrive in an environment with a lot of ambiguity, where we are growing quickly and changes happen to the product and the business on a regular basis.

You will be challenged, supported and have the opportunity to learn a lot. You will work a fast paced, autonomous environment with like minded people who are passionate about what they do.

We care deeply about helping the tech industry become a more inclusive and diverse place and we work hard to lead by example. Our workplace is dynamic, diverse and highly collaborative. Join a company with;

• 50 engineers and data scientists with plans to double the team size in the next 6 months.
• 5M duplicated products detected and merged using product image features (http://www.slideshare.net/ejlbell/fashion-productdeduplication)
• 300k online recommendation model updates per day (http://developers.lyst.com/data/2014/11/11/word-embeddings-for-fashion/)
• 72k crowd-sourced labels generated per day
• 40k product gender classifications per day via deep learning
• 500k recommended products per day
• 120 EC2, 8 RDS, 7 ElastiCache and 10 Redshift instances
• our internal analytics system collects ~100M data points/day

...and a team that… 

• ~10 deployments/day
• 40+ merged pull requests/day
• 20k lines of change/week
• Lots of open source projects - https://github.com/lyst and https://github.com/SSAW 
• Get invited to talk at great events (PyCon, Europython, PyData etc)
• feature toggling and A/B testing

• Twice monthly internal engineering meet up events
• Paid attendance at conferences
• A clothing allowance
• Internal training opportunities (want to learn Python, or improve your presentation skills?)
• Desk beer Fridays
• A well stocked kitchen and fridge
• Things that keep you happy and healthy: Yoga in the office regularly, football teams, netball teams, board game nights and burger eating clubs.