Computer Scientist (PhD) · Lossless Compression · Compressed Data Structures · Green Algorithm Engineering
🔬 Researcher at L'EMbeDS, Scuola Superiore Sant'Anna (Pisa, Italy), working at the intersection of lossless data compression, compressed data structures, and green, energy-aware algorithm engineering.
My current work designs scalable, energy-aware compression, indexing, and semantic-search techniques for large source-code archives — in collaboration with Software Heritage — grounded in a commitment to open science and open-data infrastructures.
- Compression & compressed data structures — lossless & computation-friendly compression, string indexing, stringology
- Compressed linear algebra on GPUs — grammar-compressed matrices, SpMV & PageRank kernels, memory-bounded engines
- Code archives & semantic search — source-code archives, code-to-code retrieval, digital preservation, open data
- Green algorithm engineering — energy–throughput trade-offs across the stack
Live, interactive views generated from my Wikidata record via Scholia — click any image to explore it interactively.
 Research topics |
 Co-author network |
 Collaboration geography |
- mediawiki-code2code-search — semantic code search over 1.1M+ MediaWiki snippets (Jina AI embeddings, cross-language retrieval, SWHID permalinks)
- green-compressed-storage — GiB/s Permute-Partition-Compress source-code storage on RocksDB
- g-mm-repair — memory-bounded GPU engine for right-multiplication over grammar-compressed matrices
- green-lossless-spmv — space/time/energy-optimised lossless SpMV
Maintainer of open-source tools on Wikimedia Toolforge, active member of the Indic MediaWiki Developers User Group, and a Diff author. 150+ Wikipedia articles started and structured data curated at scale on Wikidata.
📧 francesco.tosoni@santannapisa.it · 🌐 francescotosoni.it · 🦋 @ftosoni on GitLab