Thomas Reynaud Email and Phone Number

Leading a team of 10 data analysts and data engineers specialized in manufacturing and supply chain problematics.

Unbabel’s “Translation as a Service” platform allows modern enterprises to understand and be understood by their customers in dozens of languages.Powered by AI and refined by a global community of tens of thousands of human linguists, Unbabel delivers professional-grade content at the scale required by modern enterprises like Pinterest, Skyscanner, Under Armour and Rovio.Backed by Scale Venture Partners, Notion, Microsoft Ventures, Salesforce Ventures, Samsung NEXT and Y Combinator, Unbabel is accelerating the shift to a world without language barriers.

Focus on Human, we take care of dataClustree is the first SaaS artificial intelligence solution for HR decisions.We offer to HR the unique place to get justified matching at every step of employee life cycle, from recruitment to departure. Our algorithms collect, normalize and turn million worldwide career paths data into evidence-based recommendations.Our end-to-end artifical intelligence & big data technology provides call to actions and justified recommendations for successful Human Capital decisions.AwardsClustree "Cool Vendor 2016" - GartnerClustree Innovation of the year - 2016, French Ministry of Economy, Finances and Telecom ParisTechClustree HR Start-up of the year - 2015, HR Tech World CongressClustree Start-up of the year – 2015, L’Usine Nouvelle

Immersive program focused on building data science skills through project-based curriculum.• Project 1: At which subway station should one place street workers to optimize the donations for a charity gala. MTA turnstiles data, CartoDB visualisations,..• Project 2: Web scraping using BeatifulSoup and Selenium python libraries. I extracted metadata from 15k+ movies from IMDb to create an actor-actor graph. I then used features from this graph to predict Box Office revenues with a regression model.• Project 3: Built a divergent recommendation system for Yelp based on Yelp academic dataset. Given the past reviews of a specific user, recommend a totally different restaurant he could also enjoy. Benchmarked different supervised learning models such as logistic regression, SVM, neural networks and collaborative filtering.• Project 4: Analysed the semantics of Emojis on Twitter with NLP and Word2Vec. Showed that emojis can be used to enrich basic sentiment analysis models by adding a lot of different nuances. • Project 5: Building an automatic image colorization algorithm based on Convolutional Neural Networks.

End of studies internship within the operations research team at Orange Labs in Paris.Presentation: https://drcn2016.lip6.fr/slides/perrot.pdfPaper: http://ieeexplore.ieee.org/document/7470849/I worked on a combinatorial optimization problem at industrial scale aimed at finding the optimal number of controllers and their location in SDN-based architectures. We focused on finding the best trade-off between decreasing the latency, the number of controllers and the electrical consumption of the network, as well as improving its resilience.⇨ Key Accomplishments:• Literature review• Development of different mathematical models and selection of the best solving methods.• Proof of concept on smaller topologies• Heuristics and approximation algorithms development for solving the problem at scale.⇨

Second year of engineering school interniship.I worked with the "Methods and Quality team" of the biggest Cartier's high jewellery workshop in Paris.⇨ Key Accomplishments:• Transformation of projects into perennial activities and gathering of results.• Support of project managers in their missions.• Weekly meetings leading.• Organization of internal audits.• Keeping dashboards up to date.• Communication of results (newsletters, meetings reports, ...)⇨