Romain Gratier
Product Developer at Picterra
Geneva, Geneva, Switzerland
Summary
👤
Senior🎓
Top SchoolRomain Gratier is a product-focused developer and machine learning enthusiast with eight years of experience building full-stack, ML-driven geospatial products from prototype to production. Based in Geneva and trained at EPFL in civil and transportation engineering with a minor in financial engineering, he blends strong engineering foundations with data-driven product management and user advocacy. At Picterra he spans roles from backend geospatial pipelines and cloud automation to frontend WebGL mapping, while defining KPIs and translating customer feedback into UX and technical priorities. He has hands-on experience across NLP, CV, time series and MLOps from earlier roles and a brief CTO stint, which gives him both technical breadth and a founder’s perspective. Outside core engineering, he writes about his work on Medium, signaling a habit of documenting and sharing practical learnings.
8 years of coding experience
6 years of employment as a software developer
Master of Civil Engineering Transportation Engineering and a Minor in Financial Engineering, Master of Civil Engineering Transportation Engineering and a Minor in Financial Engineering at EPFL
French, English, German
Github Skills (58)
staging10
bitcoin10
cryptocurrency10
cryptography10
p2p10
extrapolation9
markov9
generative8
python-client8
bitcoin-core8
tiles8
earth-observation8
google-workspace8
mapbox-gl-js7
google-docs7
Programming languages (4)
Github contributions (5)
Deep neural networks have garnered tremendous excitement in recent years thanks to their superior learning capacity in the presence of abundant data resources. However, collecting an exhaustive dataset covering all possible scenarios is often slow, expensive, and even impractical. The goal of this project is to devise a new learning framework that can learn from a finite dataset and noisy feedback of data properties to discover novel samples of particular interest. We will design and implement algorithms to interweave emerging deep generative modeling with classical Markov decision processes. We will evaluate our method in comparison to existing approaches through extensive experiments, including but not limited to visual semantic extrapolation and natural adversarial examples in the context of autonomous vehicles.
Contributions:125 commits, 224 pushes, 1 branch in 9 months
particularextrapolationgoaladversarialbayesian-inference
RomainGratier/recipes-telegram-bot
Dec 2020 - Jan 2021
Generate and recommend recipes based on photos of your fridge
Contributions:12 commits, 1 PR, 19 pushes in 28 days
pythonrecipesfridgeimage-processingrecommend
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.