David APPADOURAI

Assistant Professor | Ph.D. Candidate
Image
David APPADOURAI

Biography

David Appadourai is a faculty member and researcher specializing in the epistemology of artificial intelligence, data science, and machine learning engineering. He is developing an epistemological approach to generative AI systems at the intersection of the philosophy of science and artificial intelligence.

His work focuses in particular on the mathematical reasoning behind models, specifically automated theorem proving (neural theorem proving), as well as on the epistemological differences between proofs generated by humans and those generated by machines.

In addition to his teaching and research activities, he designs and deploys artificial intelligence systems in production, including LLM pipelines, RAG systems, agents based on the Model Context Protocol (MCP), and full-stack applications.

He is a member of the Association for Symbolic Logic (ASL), regularly participates in seminars at the AI Observatory at Paris 1 Panthéon-Sorbonne University, and is a member of the “Virtual Reality and Emotions” (RéVE) reading group at the Sorbonne Institute for Legal Research (IRJS - UR 4150).

 

Education & Degrees

YearDegree / Institution
In progressPh.D. in Philosophy of Science—Università degli Studi di Roma “Tor Vergata.” Advisor: Professor Gabriele Pulcini. Dissertation: An epistemological approach to mathematical reasoning in AI, focusing on automated theorem proving and the epistemological differences between human proofs and machine-generated proofs.
 Master’s 2 in Logic and Philosophy of Science (LOPHISC) – Paris 1 Panthéon-Sorbonne University. Master’s 2 thesis: “Evaluation of Computational Creativity for Symbolic Music Generation Models,” advised by Professor Marco Panza, research director at the CNRS (16/20). Master’s Thesis (M1): “A Study of Set Theory Based on Attempts to Resolve the Continuum Hypothesis,” supervised by Jean Fichot, associate professor at the IHPST (17/20).
 Bachelor's degree in Mathematics and Computer Science Applied to the Humanities and Social Sciences (MIASHS), with a concentration in economics—University of Paris (now Université Paris Cité).
 Bachelor's Degree in Philosophy, Logic and Philosophy Track - Paris 1 Panthéon-Sorbonne University.

Professional experience

Teaching and Academic Responsibilities

YearFunction
2025–presentaivancity School for Technology, Business & Society (tenured since September 2026) - Epistemology of artificial intelligence: the limitations of current models; AI ethics and the robustness of learning systems; model interpretability and AI safety; equity and Fair ML; Linux environment (MSc in Data Engineering & Cloud Computing); Coding for AI & Data Science; Model Context Protocol (MCP) workshop—agent architectures and integrations. Courses taught in French and English (Grande École Program and MSc).
2026–presentTeaching Assistant - Paris 1 Panthéon-Sorbonne University. Computer Science Department: Programming, Algorithms, and Logic.
2024–2025Teaching Assistant - University of Tours. Data Science: Statistics and Data Mining, Master 2 in Digital Humanities.
2024Instructor - École Pratique des Hautes Études (EPHE-PSL). Bachelor’s Degree in Sciences of a Sustainable World: Introduction to Music Programming (Sonic Pi).

Responsibilities at aivancity

  • Program Director for the MSc in Applied Artificial Intelligence Online—responsible for program design, coordinating faculty, and developing the seven course units.
  • Co-supervision of PGE4/PGE5 research projects on the explainability of AI agents.

Entrepreneurship and Engineering

YearFunction
2021–presentTechnical Consultant - Parisian Spirit / TV STORE, Cachan. NDI video streaming network deployed across 80 Raspberry Pi nodes (photo booth kiosks, video walls, interactive AI installations) for a European specialist in audiovisual event solutions.
2020–presentFounding President and Technical Director—Vidyalaya Association, Paris. Teaching French as a foreign language (DELF preparation) to migrants and providing academic support in working-class neighborhoods. Design of the entire digital infrastructure: Vidyia (conversational AI tutor—RAG based on the national curriculum, visual essay grading, text-to-speech), Next.js student portal, and DELF worksheet and test generator.
2023–presentCo-founder - Maison ANTIK. Premium rum brand: website Next.js/Vercel, automated SEO pipeline, and an autonomous operational agent based on an LLM, handling customer relations, billing, and business process automation.
2021–presentFounder and Developer - Aleph Music Publishing (SAS), Villejuif. Music publishing: automation of catalog publishing tasks and an app that enables independent artists to act as their own publishers.
2015–2024Translation Assistant and Technical Developer - T.I.A.S.C.I, Paris. Translation of official documents for a certified translation agency; design of a machine translation assistant combining computer vision and an LLM API pipeline.

Publications

Publication

  • 2025 - “The Avatar: An Epistemological Sketch,” in “Virtual Reality and Emotions” (RéVE) Research Group, summary document, Paris 1 Panthéon-Sorbonne University. HAL: hal-05081325.

Manuscripts in Progress

  • A bidirectionality criterion for machine-verified proofs (philosophy of mathematical practice) and two related empirical studies auditing Lean proofs accepted by the kernel and generated by state-of-the-art neural demonstrators.

Scientific Communications

  • July 2026 - Logic Colloquium 2026 (ASL European Summer Meeting, FLoC), Swansea (United Kingdom): “Localizing Overgeneration: A Three-Strata Decomposition of Formal Proof.” Paper presented.
  • September 2026 - Human-Oriented Automated Reasoning (HAR 2026), Paris. Paper presented.
  • November 2026 - Trends in Logic XXVI, Sendai (Japan). Paper accepted.
  • 2026 - Conference on Artificial Intelligence and Theorem Proving (AITP 2026). Extended abstract submitted.
  • September 2024 - Paris 1 Panthéon-Sorbonne University, conference on “Synesthetic, Mathematical, and Memory Transfers”: “Transfer Learning in Music Generation Based on Diffusion Models.”
  • July 2024 - Una Europa Summer School on Artificial Intelligence & Creativity, CEUB, University of Bologna.

Translation

  • 2022–2023 — French translation of *Introduction to Proof Theory* by Professor Paolo Mancosu (University of California, Berkeley), 588 pages — IHPST, Paris.

See also