Freelancer profile translated to English.

Description

AI Research Engineer, from the joint Mathematical Modeling program between École Polytechnique and Sorbonne Université. I support quantitative funds, research labs, and engineering teams in the architecture and deployment of cutting-edge algorithmic systems.

Shunning a generalist approach, my expertise is strictly focused on mathematical rigor, optimal control theory, and software performance in complex environments (Python, C++).

My technical areas of intervention include:

-Generative Artificial Intelligence: Fundamental research and implementation of Diffusion Models (DDPMs), analysis of convergence rates, and advanced RAG system architecture.

-Quantitative Engineering: Development of rigorous statistical evaluation engines, probabilistic scoring, and implementation of low-latency algorithms.

-From Research to Production: Reproduction of state-of-the-art (SOTA) research papers and delivery of reliable engineering artifacts (evaluation pipelines, benchmarks) to de-risk model deployment.

Examples of recent implementations:

Founding EAVAE Labs, an AI engineering firm specializing in creating evaluation infrastructures for the production deployment of complex models.

I do not engage in classic web development or basic API integration. If you are looking for a level of academic rigor coupled with industrial execution capability for your artificial intelligence architectures, contact me for an audit of your systems.

Languages

French
Native or bilingual
English
Native or bilingual
Arabic
Native or bilingual
Italian
Native or bilingual
Russian
Basic

Workplace preferences

Can work on-site

Paris (up to 50km)

Université Paris 1 Panthéon-Sorbonne
AI Researcher
PUBLIC SECTOR
April 2026 - Today (2 months)
12 Pl. du Panthéon, 75005 Paris, France
SAMM Lab - Study of the impact of regularization on score learning in diffusion-based generative models.
Machine Learning Neural Networks Data Science Computer Vision
moltsmarket
Co-Founder
January 2026 - Today (5 months)
San Francisco, CA, USA
Developing moltsmarket, a forecasting and reputation platform for human analysts and AI agents leveraging React and Next.js.
Building a global proving ground designed to counter market noise and AI hallucinations by providing verifiable, mathematically
scored track records. Implemented an immutable ledger system for timestamping predictions and designed the frontend architecture
to support independent analysts, institutional allocators, and AI agent developers.
- Architected and implemented an immutable claim system, ensuring all forecasts are permanently timestamped, locked, and
bound to reality.
- Integrating a rigorous statistical scoring engine utilizing Brier scores and Kelly logs to evaluate and grade user and AI agent
calibration and accuracy.
- Developing syndication features that allow top-calibrated forecasters to monetize their reputation through premium thesis feeds.
- Designing a responsive web application utilizing Next.js for a seamless user experience.
latentQ
Founding AI Engineer
December 2025 - February 2026 (2 months)
Architected and developed the core intelligence engine for a comprehensive FAANG and quantitative finance interview simulation
platform. Designed AI-driven assessment systems to automatically evaluate, score, and rank candidates across Software
Engineering (SWE), Machine Learning Engineering (MLE), and Quant roles. Built scalable infrastructure to support full interview
loops, including automated Online Assessments (OAs), technical problem-solving, and behavioral rounds. Collaborated on the end-
to-end product lifecycle, from designing the problem evaluation logic to implementing the user-facing simulator.
- Engineered an automated evaluation pipeline capable of assessing complex ML models (e.g., next-day return micro-models)
using metrics like MSE and directional accuracy against hidden test data.
- Developed a dynamic ranking algorithm that accurately calculates live percentiles for candidates against real-world industry
benchmarks from top-tier firms like Google, Jane Street, and Citadel.
- Built targeted assessment pathways for highly specialized roles, including ML-driven Quant Research, Statistical Arbitrage, and
Market Making Research.
- Played a pivotal role in scaling the platform to support hundreds of coding problems and live interview simulations with real-time
feedback.