Roberto Trani

Software Engineer @seltz_ai
PhD in Computer Science @Unipisa

[Antonio Mallia @antonio_mallia]

Today we're announcing our $12.5M seed, led by B Capital and Speedinvest.

In 8 months, a small team built a news index that's faster and more accurate than any search API we tested, and that hands agents full, fresh documents instead of snippets.

[Seltz @seltz_ai]

Building Seltz's Cloud Infrastructure from the Ground Up 🚀

Excited to share what our team has been cooking up at Seltz. @aljohri has been absolutely crushing it on our cloud infrastructure, architecting a modern Internal Developer Platform that's setting us up for serious scale

[Seltz @seltz_ai]

🚀 Built from scratch. Because scale doesn't tolerate shortcuts.

When you crawl hundreds of millions of pages per day, "off-the-shelf" quietly stops working.

So we didn't use it.

[Seltz @seltz_ai]

🚀 A learned lexical model for Web search — and why it matters

At Seltz, we’re building search end-to-end, models included.

For Web retrieval, we wanted something very specific:
a model that beats BM25 on Web benchmarks, while keeping the core strength of lexical retrieval

[Seltz @seltz_ai]

At Seltz, keeping everything under control means measuring everything that matters in search.

Since joining the team, @EliasBassani built a monitoring & eval framework to:
• Measure retrieval quality
• Measure agent task success
• Track retrieval + generation latency

[Patrick Moelleken 🇺🇦 @PMoelleken]

.@ZelenskyyUa's tv address to the Russian (!) people might be the most moving speech that I've ever seen in my entire life. The whole world needs to see, understand and share this crucial Ukrainian message.
#StandWithUkraine #Ukraine #Україна #Russia #Россия

[takuya-a @takuya_b]

I introduced PTHash algorithm paper at the in-house paper reading session. It was not easy to explain it and its background within one hour but we enjoyed that time. https://t.co/f1UaYS8H4C

[Franco Maria Nardini @fmnardini]

"Fast Filtering of Search Results Sorted by Attribute" ACM TOIS. @Roberto_Trani @DivBy now online: https://t.co/G2qOZVWFkb

[Giulio Ermanno Pibiri @giulio_pibiri]

Just in case you missed my most recent project,
with @Roberto_Trani. Minimal Perfect Hashing this time ;)
Papers: https://t.co/RfFuEloUFt
https://t.co/0N9FrCXH40
Code: https://t.co/WCkfznQFZY

[HPC Lab @ ISTI-CNR @HPCLab]

Full paper #SIGIR2021:
PTHash: Revisiting FCH Minimal Perfect Hashing.
Giulio Ermanno Pibiri, @Roberto_Trani
@SIGIRConf

[HPC Lab @ ISTI-CNR @HPCLab]

"Fast Filtering of Search Results Sorted by Attribute" @fmnardini @Roberto_Trani @DivBy Journal Paper accepted on @acmtois Transactions on Information Systems (TOIS)

[𝕐 @nomad421]

In 2016/7 BBhash created an absolute revolution in usability at scale of minimal perfect hashing. 4 years later, another very interesting approach looks to have been introduced https://t.co/tkm9Bfm2Ur! Any thoughts on this approach @RayanChikhi / @BQPMalfoy?

[ECIR 2021 @2021Ecir]

Full program of #ECIR2021 finally published!
https://t.co/AvFJY9xDGu
@2021Ecir #ECIR
@HPCLab

[Roberto Trani @Roberto_Trani]

'Fast Approximate Filtering of Search Results Sorted by Attribute' has been accepted as a full paper at #SIGIR2019 w/ @DivBy and @fmnardini.

[Roberto Trani @Roberto_Trani]

@RoombaSoul and what do you do tomorrow?

[Roberto Trani @Roberto_Trani]

@RoombaSoul What is the prognostic for the match of tomorrow?

[Franco Maria Nardini @fmnardini]

#SIGIR2016 conference opening by @raffaele_perego!

[Roberto Trani @Roberto_Trani]

UnRisk Insight: Fast Functional Goats, Lions and Wolves http://t.co/QJ60S4aYO5

[Feross @feross]

Facebook's git repo is 54 GB.