Skip to content

About

Vincenzo Imperati

I build and operate data infrastructure: systems that acquire data other people find hard to get, keep it correct, and expose it so it can be used.

That work spans BigBrotr, an OpenSats-funded network observatory I run on hardware I maintain; the analytical data layer of a live game; and doctoral work in large-scale web measurement.

The research is published at USENIX Security and IEEE Access. Most of the engineering is open source, including contributions to infrastructure I use.

Experience

Aug 2025 – Present · Zurich

Software Engineer (grant-funded) · OpenSats · BigBrotr

  • Won competitive OpenSats funding to take a research idea to production, and designed, built and now operate the resulting platform single-handed.
  • Sustained ingestion and archiving of 250M+ signed events from 5,000+ independent relays, covering 25M+ public keys, by coordinating eight asynchronous Python services through durable PostgreSQL state with replayable checkpoints.
  • Held a sustained ingest rate of 2,000+ events per second, peaking at 3,000, against a live network of independent relays.
  • Made the pipeline recoverable under partial failure by routing coordination through durable state instead of direct RPC, which cost extra write volume and bought independent restarts and replayable checkpoints.
  • Reworked the storage layer for analytics workloads with 16-way hash partitioning, LZ4 compression, redesigned indexes and materialized views.
  • Kept the public archive free of duplicates at ingest time by combining in-memory deduplication with a content-addressed SHA-256 composite key, so re-ingesting the same event is a no-op rather than a reconciliation job.
  • Built one schema-introspecting query layer that both the read-only FastAPI service and the Nostr data-vending-machine serve from, so filter, sort and pagination semantics are defined once rather than per surface.
Read the case study →

Nov 2024 – Mar 2025 · Rome

Software Engineer · Freelance

  • Delivered the field application for Gen4Olive, an EU Horizon 2020 research project, giving agronomists access to 600+ olive cultivars with camera capture, image preprocessing and model-assisted cultivar and disease recognition.
  • Replaced an Italian SME's paper-based HR records with a Next.js/Firebase operations system, turning record retrieval into a single query, with searchable profiles, document storage, role-based permissions and per-trade competence ratings.

Sep 2022 – Aug 2024 · Remote

Data Engineer · Cantina Royale

  • Unified four heterogeneous sources – on-chain transactions, gameplay telemetry, asset metadata and market activity – into the single analytical layer the live game's reporting ran on.
  • Caught and escalated a reward-farming exploit by reconciling on-chain reward flows against off-chain gameplay outcomes, which no single data source could have revealed.
  • Shipped the player-facing analytics platform on a server-side SQLite layer, cutting client payload size with compact paginated responses.
Read the case study →

Apr 2021 – Jul 2023 · Rome

Teaching Assistant · Sapienza University of Rome

  • Supported 200+ students across four Computer Science courses (programming, algorithms, object-oriented programming, algorithm design) through weekly sessions, exercise review and exam preparation.

Education

Nov 2023 – Oct 2026

PhD in Computer Science

Sapienza University of Rome

Data infrastructure, distributed systems, and trust and safety in large deployed open systems.

Sep 2025 – Aug 2026

Visiting Researcher · ETH Zurich · University of Zurich

Empirical measurement of deployed distributed systems and open protocols, including structural-centralization risk across the topology of 5,000+ relays, with reproducible collection and analysis.

Sep 2021 – Oct 2023

MSc in Computer Science

Sapienza University of Rome

110/110 with honours, Honors Program excellence track (roughly 3 of 150). Thesis on conspiracy-channel detection and monetization at scale, the basis for the USENIX Security 2025 paper.

Sep 2018 – Jul 2021

BSc in Computer Science

Sapienza University of Rome

110/110 with honours, Honors Program excellence track (roughly 6 of 300). Thesis on reinforcement learning for home-energy management, smart-grid optimization and EV charging.

2025

The Conspiracy Money Machine: Uncovering Telegram's Conspiracy Channels and their Profit Model

USENIX Security 2025

Vincenzo Imperati, Massimo La Morgia, Alessandro Mei, Alberto Maria Mongardini, Francesco Sassi

2025

Hard to See, Harder to Block: Exploring the Challenges in Email Tracking Pixels Detection

IEEE Access

Federico Cernera, Vincenzo Imperati, Massimo La Morgia, Alessandro Mei, Alberto Maria Mongardini, Eugenio Nerio Nemmi, Luciano Ortenzi, Francesco Sassi

Open-source contributions

Eighteen merged pull requests across eight repositories I do not own.

Protocol specifications

Four merged changes to the Nostr NIPs, the protocol specification itself: normalising `d`-tag handling for non-Latin scripts in NIP-54, migrating that NIP from Asciidoc to Djot, clarifying `limit` filter behaviour in event queries, and a pass over broken links across several NIPs. Two further proposals extending NIP-66 with DNS, SSL and geo checks are open.

nostr-protocol/nips →

Infrastructure and storage

Two merged fixes to fiatjaf/eventstore, the collection of storage backends for Nostr events, including a correctness fix where tag values were not prefixed with their tag key and so produced wrong results for NIP-01 filters. On nopayloaddb, the HEP Software Foundation reference conditions database used by sPHENIX and Belle II, a correctness fix to the interval-of-validity computation is merged and further work on query efficiency and error handling is open.

fiatjaf/eventstore →BNLNPPS/nopayloaddb →

Security and data protection

Seven merged pull requests to bio-block, a health-data platform: sanitising formula injection in anonymised exports, closing an ownership-hijack path in the document-storage contract, adding PHI-detection tests to the anonymisation pipeline, and adding input validation to the search API.

healthyinc/bio-block →

Research tooling

Merged contributions to three repositories of the Sapienza systems lab, including the Telegram dataset tooling that underpins the measurement work.

SystemsLab-Sapienza/TGDataset →

Selected projects

All projects →

Skills

Languages

PythonSQLTypeScriptC++CBashJavaScript

Data & storage

PostgreSQLSQLiteDuckDBRedisLMDBETL pipelinesschema designpartitioningindexingquery optimisationmaterialized viewsdeduplicationprovenance

Backend & distributed systems

asyncioFastAPIDjango REST FrameworkasyncpgaiohttpREST API designWebSocketsmessage queuesCelery

Acquisition

web crawling and scraping at scaleHTML/JSON extractionrate-limit and anti-bot handlingnormalisationlarge-scale data processing

Infrastructure & operations

DockerDocker ComposeLinuxProxmoxPgBouncerPrometheusGrafanaAlertmanagerGitHub ActionsCI/CD

Quality

pytestmypyruffcode reviewautomated testing

AI tooling

Model Context Protocol (MCP) serverssemantic search and vector storesagent-oriented tooling

Learning now

GoRustRocksDBClickHouse

Awards & recognition

  • 2026Built with Claude Code programSelected as one of 500 builders from more than 20,000 applicants.
  • 2025OpenSats GrantCompetitive open-source funding for BigBrotr.
  • 2025Outstanding University StudentSapienza University of Rome.
  • 2023–202517 prizes across 10 international hackathonsLeading teams of 3–5 from idea to working prototype in 24–48 hours.
  • 2023MSc Honors ProgramExcellence track, roughly 3 of 150 students.
  • 2021BSc Honors ProgramExcellence track, roughly 6 of 300 students.