Pascal DUVAL

Déploiement de pipeline RAG basé sur LangChain/M Évaluation de performances LLM (OpenAI/Mistral) Études exploratoires (ACP, clustering, NLP) Déploiement d’API IA sécurisées et scalables

Formateur référent en Tokenomics

currently working on a Dao (called ZEST) granting cultural projets both with private and institutionnal funds. Assembling building blocks (quadratic vote and funding, bonding curve) in solidity ethereum as a proof of concept ; then migrating to substrate (polkadot) and/or optimitic ethereum.

* In charge of the design of a web platform offering on demand visual recognition services (video, pictures) on a vast corpus (french press). - Benchmark of the best solutions on the market - Set up with a spinoff (of a renowned center of research) of a product image search and publication stats tools of these images. * In charge of the integration of Named Entites Engines in the flow of press news and in charge of the interfaces enrichment. The process is this : Input -> A continuous stream (24h/24h) annotated with Named Entities [People, Companies, Cities and Relationships]. Annotation is central for structuring the indexing of these articles (for example with press events and companies informations) . Output -> The enriched articles (xml) are sent to applications or search engines. - Knowledge of integration problems (including Web Services) and how to handle several extraction tools for information retrieval. - Knowledge of several Entitites extractor (Open or not) and good (best, I hope) practices issues regarding annotation (especially in terms of precision / recall). * Proposal of an inductive learning strategy to deal with the automaton of the indexing process #Understanding Domain, Prior Knowledge and Goals #Data Integration, cleaning, pre-processing #Learning Models (Semi-Supervised ones like Bayes, SVM,..) #Interpreting Results #Loop

This PhD in Philosophy is entitled: Stanley Cavell's perfectionism – From the authorization of oneself towards a politic of interpretation. The main purpose of this PhD lies in the similarities between Nietzsche and Emerson that Stanley Cavell has revealed. A whole “politics of the interpretation” pressing on the name of Nietzsche has to be revised. The thought of Nietzsche is not to be viewed as an enemy of freedom. Would it serve in the same way as Emerson a democracy which would in return be honoured by it? Is this question simply possible given the hermeneutic turn and the horizon of a closing of metaphysics defended by an influential part of the contemporary philosophy (which sometimes claims to refer to Nietzsche) here and there given the dominance of the most blind political theory ? There is, however, a hope of a new step in the arisen of the most bursting return of America to its emersonian voice: the election of B. H. Obama. A “new difference”, and a success of perfectionism. But to make the other half of the way, we had to recount the critical enterprise of Stanley Cavell in its «modernist» moment : the discovery of a certain form of authority in philosophy or the «overtaking of the philosophical theory». The essence of what is a «text», its powers and its places are changed according to a new auctoriality which finds its immediate data in a relieved culture.

Statistician Assistant