Jair Wuilloud

Jair Wuilloud

Data Scientist / Engineer / Researcher

I see myself as an entrepreneurial problem solver and a team player.

PDF JSON
Location
27A Sylvester Road, N2 8HN, East Finchley, London, United Kingdom
Email
GitHub
gilgameshjw
LinkedIn
Jair Wuilloud

Experience

– present

Consultant at Jair ltd

Consulting in data science for more than 30 entities, including well established companies and some of the most ambitious startups. Some of the work is described below or in Projects.

–

Lead AI Engineer at JUA AI

I was invited by JUA.AI to build advanced weather models using deep learning. Starting from a blue sky project, within 6 months, I could identify, test, train and integrate models and data architecture components that became the core of currently the best weather model predictions on the market.

–

AI Engineer at Interscript

At Interscript, I developed a solution to deliver transliteration with generative models. First, diacritization models were trained in arabic and hebrew, resp. reproducing and beating existing benchmarks. Next, I have designed a new approach to transliteration allowing language experts to design the necessary complex transliteration rules. Data and models are then built and trained via a semi automatised strategy that was presented on a conference to experts.

–

Co-Founder at Neurotrophic Labs

Together with a MD as a cofounder, we have developed an app for the US market allowing practitioner to easily model patient states and diseases and retrieve statistical evidences in real time via fuzzy search throughout the US healthcare dataset. The app’s focus on speed allows running medical scenarios and test hypotheses in real time.

–

Co-Founder at Plex AI

Plex AI vision: interaction with websites via language. My main contribution was building of an iOS app, the agent, the design of a neuro inspired architecture to compress the conversational logic and to run successful demos to potential customers.

–

Lead Data Scientist at MediaPlayers

I was invited at Mediaplayers to develop their data strategy. I built a data pipeline processing within 3 hours ~1 billion of daily behavioural data points.

–

Senior Data Scientist at Faxi (now kinto join)

For faxi, I built new GPS processing and carpooling recommendation algorithms. Thanks to an app that I created, communities can now easily visualise the carpoolings and their benefits in term of money saved and CO2 emissions.

–

Senior Data Scientist at Constellation AI

I joined the company to build new language models using advantages of sparse representations. After 3 months, I developed a working model to encode and compress text that became the semantic engine after 6 months. This model was extended and tested on several business cases. The model proved to exceed some existing benchmarks, which we have documented in a paper.

–

Founder at TaxiQ

At TaxiQ we created an app based on collaborative system helping cab drivers in UK to navigate within the city by solving the problem of evaluating the latest waiting times on the taxi ranks. As a CEO, I have managed to put together a team of up to 8 people and to build a MVP (web apps) with a small budget. Along with R&D and development, I developed the business plan and was pitching to potential investors and competitions.

–

Model Validation at Deutsche Bank

At DB, I worked on the validation of financial models and implementations on the C++ pricing libraries.

–

Postdoctoral Researcher at Institute for Theoretical Physics

I was working on QCD and Supersymmetry simulations. Simulations allow to run virtual experiments and perform approximations necessary to do predictions and test the standard model. I have specialised in some technical aspects of DIrac’s operator determinant evaluation.

Education

–

Phd in Numerical Particle Physics from University of Muenster (WWU) with Summa cum Laude

–

Master in Mathematical Physics from University of Geneva with 5.25/6

Publications

Jua Vilhelm: A novel end-to-end AI system for global weather prediction

High Dimensional Computing for Public Safety

Carpooling vs Public Transport: Impact on COVID spread

Fast and light GPS processing via Coarse Coding

IVA, a Chatbot coordinated via Sparse Distributed Memories

Superpositions with Sparse Distributed Memories, High Dimensional and Dense Vectors

System and method for encoding multiple Information sources including Human Language

Past Research in Particle Physics

Theoretical particle Physics.

Languages

French
Fluency: Native speaker
English
Fluency: Full professional proficiency
German
Fluency: Full professional proficiency
Italian
Fluency: Full professional proficiency
Portuguese
Fluency: Elementary proficiency

Skills

Programming
Level: 2.5/3
Keywords:
  • Julia
  • Python
  • C++
  • clojure(script)
ML & Deep Learning
Level: 2.5/3
Keywords:
  • Sklearn, Pytorch, Keras, Flux.jl
AI
Level: 2.5/3
Keywords:
  • Generative Models
  • Brain inspired Computing
  • Reinforcement Learning
  • Conversational Agent
  • Collaborative Systems
  • Simulations
Data Processing/Modelling
Level: 2.5
Keywords:
  • Spark, Hive
  • NLP, GPS, Behavioural data
Maths & Physics
Level: 2.5/3
Keywords:
  • Statistics, Inference Learning
  • Financial Maths
  • Optimisation
DevOps
Level: 2.0/3
Keywords:
  • AWS, Linode, Azure
  • docker, git
  • Bash, Web Protocols

Interests

Classical Singing, Strategy Games, Sports
Keywords: