Publications
In Preparation / Under Submission
- GANTIS: Graph-based Approximate Nearest Neighbor Search for Varying Length Time Series (2025)
- ELPIS+: Scalable Similarity Search for High Throughput and Latency (2025)
- GANNS on Steroids: Optimized Graph-Based Approximate Nearest Neighbor Search (2025)
Work Experience
- Postdoctoral Fellow, CNRS & UM6P (2025–Present) – Conducting research on advanced AI-driven retrieval systems and scalable similarity search at Paris Cergy University.
- Applied Scientist Intern, Amazon (2024) – Designed and implemented optimization techniques for large-scale vector search systems, enhancing retrieval speed and scalability for next-generation AI applications.
- Teaching Assistant, Université Paris-Cité (2022–2023) – Delivered Python programming labs for undergraduate students.
- Teaching Assistant, UM6P (2022–2023) – Supported Database Management lectures and labs for engineering students.
- Machine Learning Engineer, Farasha Systems (2019) – Developed ML models for AI-assisted predictive maintenance at Noor I CSP solar plant, Ouarzazate, Morocco.
Talks and Presentations
- University Paris Dauphine – ENS – Gave a lecture on vector databases and graph-based vector search for M2 students in the Data Management course.
- Amazon AGI – Delivered a talk on scalable similarity search and graph-based indexing methods for large-scale vector data.
- VLDB 2024 – Presented PhD workshop paper on billion-scale vector search and hybrid indexing techniques.
- UM6P Science Week 2024 – Invited speaker on AI-driven retrieval systems and efficient similarity search techniques.
- Meta FAISS 2024 – Gave a talk on graph-based similarity search and introduced ELPIS for efficient vector indexing.
- VLDB 2023 – Presented ELPIS at the main conference session on similarity search for high-dimensional vectors.
- DiNo Seminars – Regular contributor and presenter on scalable data systems and graph-based retrieval methods.
- UM6P-CS Seminars – Gave talks on similarity search, indexing algorithms, and AI applications in large-scale data environments.