Aditya Parameswaran
adityagp.bsky.social
Aditya Parameswaran
@adityagp.bsky.social
Prof @ UC Berkeley. Codirector @ EPIC data lab. Cofounder @ Ponder (Acq. Snowflake). Interested in data, systems, and people. More at: adityagp.net
Reposted by Aditya Parameswaran
Beautiful nostalgic piece by @adityagp.bsky.social! Don't miss the end for some nice reflections. Aditya's just getting started! :)
October 1, 2025 at 10:26 PM
Blog Post: Looking back on the first decade as faculty (2014-2024).

I list my favorite papers from the decade, why I enjoyed working on them, and provide backstory and reflection.

data-people-group.github.io/blogs/2025/0...
The First Decade as Faculty - Looking Back
A research blog on next-generation data infrastructure that is usable, intelligent, and efficient
data-people-group.github.io
October 1, 2025 at 8:29 PM
Reposted by Aditya Parameswaran
🧵 A new public database, the Police Records Access Project, now makes 1.5 million pages of public records on police use of force and misconduct in California searchable for the first time.

The project was led by @berkeleyjournalism.bsky.social, @ucbids.bsky.social, and Stanford's Big Local News.
August 4, 2025 at 3:34 PM
Reposted by Aditya Parameswaran
“Inside California’s new police records database” - Spend a few minutes watching an interview with Cheryl Phillips (@cephillips.bsky.social) on KTVU Fox 2 (@ktvufox2.bsky.social) to hear more about the new database.
www.ktvu.com/video/1686149

#AI #UCBerkeley #Stanford #localnews
Inside California’s new police records database
Cheryl Phillips, founder and co-Director of Big Local News at Stanford University, gives insight on new database giving Californians access to nearly one-point-five million pages of internal police re...
www.ktvu.com
August 6, 2025 at 4:49 PM
Reposted by Aditya Parameswaran
Today, we celebrate the incredible work on the Police Records Access Project – a first-of-its-kind database in the nation. It took years of work by a multidisciplinary team of journalists, data scientists, lawyers and civil rights advocates.

Learn more: bit.ly/4m5uDlu
August 5, 2025 at 12:10 AM
Reposted by Aditya Parameswaran
Reposted by Aditya Parameswaran
We're hiring! Working with @mikeolson.mastodon.social.ap.brid.gy and the rest of our @aimatx.bsky.social team has been wonderful, but we have a lot more to accomplish.

If you're interested AI, material science (theory + exp) and user-facing tools, join us!

www.linkedin.com/posts/fperez...
www.linkedin.com
February 25, 2025 at 4:56 PM
My periodic reminder to my CMU friends to get someone with any inkling of UI design to fix their letter upload portal. I don't know what to pick between "Submit" and "Upload Recommendation", and the dropdown forces me to make a choice even if I don't want to. cc @andypavlo.bsky.social @domoritz.de
December 7, 2024 at 7:21 PM
Reposted by Aditya Parameswaran
If you're at The Curve Conference this weekend, come dive into what the future of applied evals looks like! I'll be demo-ing EvalForge, an implementation of the "who validates the validators" paper with @weightsbiases.bsky.social

Big thanks to @capetorch.bsky.social and Anish for their work here
November 22, 2024 at 5:29 PM
Reposted by Aditya Parameswaran
Vol:17 No:12 → Dealing with Acronyms, Abbreviations, and Typos in Real-World Entity Matching
👥 Authors: Joshua Wu, Dixin Tang, Nithin V Chalapathi, Tristan Chambers, Julie Ciccolini, Cheryl Phillips, Lisa Pickoff-White, Aditya ...
📄 PDF: https://www.vldb.org/pvldb/vol17/p4104-tang.pdf
November 20, 2024 at 9:00 PM
Reposted by Aditya Parameswaran
First post here! I'm currently researching how LLMs can be grounded in tables, bc they're a goldmine for fresh domain data.

We've now introduced the 🎯TARGET benchmark for evaluating table retrieval in RAG pipelines, e.g. QA/factver/text2sql: target-benchmark.github.io

Thinking BM25? Think twice..🧵
November 4, 2024 at 5:29 PM
Reposted by Aditya Parameswaran
🎦 Watch the recording of @madelonhulsebos.bsky.social seminar at BIDS!
"There is actually a lot of structured data available on the web.... But really, the task here is that we should be able to retrieve that easily."
#datascience #structureddata #dataretrieval

bids.berkeley.edu/news/madelon...
October 30, 2024 at 5:28 PM