Tech At Bloomberg (@TechAtBloomberg) header
Tech At Bloomberg (@TechAtBloomberg) profile picture

We are engineers and designers, builders and makers delivering unrivaled products and services every day. We have a lot to share and a lot to learn. Join us.

Joined August 2014
Photo by Tech At Bloomberg (@TechAtBloomberg)Photo by Tech At Bloomberg (@TechAtBloomberg)Photo by Tech At Bloomberg (@TechAtBloomberg)Photo by Tech At Bloomberg (@TechAtBloomberg)Photo by Tech At Bloomberg (@TechAtBloomberg)Photo by Tech At Bloomberg (@TechAtBloomberg)Photo by Tech At Bloomberg (@TechAtBloomberg)Photo by Tech At Bloomberg (@TechAtBloomberg)Photo by Tech At Bloomberg (@TechAtBloomberg)Photo by Tech At Bloomberg (@TechAtBloomberg)
In today’s #QuantMinds talk “Beyond comparable company analysis: Forecasting financial variables for private firms under incomplete information” (14:00 GMT), Svetlana Borovkova presents a potential framework for modeling financial risk for private firms bloom.bg/48k6liu
1
206
In today’s #QuantMinds talk (2 PM GMT), @Bloomberg's Global Head of Quant Analytics Fabio Mercurio presents autoencoder-based forward-rate models that are arbitrage free & can calibrate historical curve shapes, as well as implied #volatility bloom.bg/4zQCiLd #QuantFinance
226
In today’s @QuantMinds 2026 plenary (9:20 AM GMT), @Bloomberg’s Bruno Dupire, Head of Quantitative Research, presents “Computing option prices when no options are traded.” bloom.bg/4AzRZq2 #QuantMinds #QuantFinance
245
In today’s #QuantMinds talk (3:15 PM GMT), Florian Bourgey presents QRH+, a simple extension of the Quadratic Rough Heston model with a one-sided, path-dependent variance boost to capture the extreme skew & put-wing curvature of short-dated SPX smiles bloom.bg/3WDaDyV
271
In Poster Session 3 at #ICAIF2026 today (15:30 CET), Ashwin Bhola, Saher Esmeir, Georgios Giannakoulias & Serge Yaskevich present a poster for "Beyond Holdout Error: A Layered Framework for Evaluating Estimators under Selective Labels" bloom.bg/3UTjQCr #AIinFinance (1/4)
1
235
Standard holdout accuracy can be a trap under selective labels. Our team's new layered framework adds distributional diagnostics across the full universe and downstream decision-stability tests - revealing estimator rankings can reverse across layers. #ICAIF2026 (3/4)
1
89
Read their whole "Beyond Holdout Error: A Layered Framework for Evaluating Estimators under Selective Labels" paper at bloom.bg/491fZ9P #ICAIF2026 #AIinFinance (4/4)
75
In Oral Session 13 at #ICAIF2026 today (14:30 CET), Gary Kazantsev presents "Adaptive Financial Time Series Tokenization with Bilevel Optimization" bloom.bg/4sXwLQp #QuantFinance #AIinFinance (1/4)
1
2
323
In bilevel financial tokenization, a leader network equalizes volatility across tokens while a follower keeps token durations predictable - competing objectives that can't be solved independently. Results: better tail-risk estimates + more stable sequence lengths #ICAIF2026 (3/4)
1
94
Read the whole "Adaptive Financial Time Series Tokenization with Bilevel Optimization" paper, co-authored by Jason Bohne, Haohan Xu, Pawel Polak, Gary Kazantsev, David Rosenberg, Ajay Alva and James Jarvis at bloom.bg/4t3ypjx #ICAIF2026 #QuantFinance #AIinFinance (4/4)
122
In this afternoon's #ACMCCS2026 talk (13:30 CET), Hamed Ghavamnia tackles a subtle CHERI problem: how capabilities can leak across supposedly isolated compartments via shared heap memory + how to fix it using type-directed capability sealing (with minimal overhead) #Cybersecurity
1
2
5
445
The paper, co-authored by Caltech Ph.D. student @MauriceHanisch & Bloomberg's @jvanegue, also found & responsibly disclosed real allocator bugs - two jemalloc exploits called Shadow Allocation & Unpainted Capability Windows, now patched upstream - then built a fix #Cybersecurity
1
1
126
Read their whole "Mind the Heap: Preventing Capability LeakageAcross CHERI Compartments" paper at bloom.bg/4iUul0h #ACMCCS2026 #Cybersecurity
1
152
In Oral Session 11 at #ICAIF2026 today (10:00 CET), Ahsan Alvi presents "No Universal Comparable: Adaptive Peer Selection for Private-Company Valuation" bloom.bg/3XsWYuo #AIinFinance #PrivateMarkets (1/4)
1
1
337
Instead of applying a fixed peer-construction rule, a learned selector chooses among methods per target firm using pre-IPO characteristics & #valuation diagnostics, and validating them against post-IPO valuation - reducing errors while preserving traceability. #ICAIF2026 (3/4)
1
86
Read "No Universal Comparable: Adaptive Peer Selection for Private-Company Valuation," a paper co-authored by Ahsan Alvi, Thomas Fonlladosa, Ching-Yu Lin, Ashwin Bhola, Chelvi Paramanathan, Duc Nguyen and Saher Esmeir at bloom.bg/4kiGuia #ICAIF2026 #AIinFinance (4/4)
1
83