profile picture

David Herel

Co-Founder & Tech Lead at BottleCapAI
Building next-gen AI that runs faster, cheaper, and smarter.
Previously: Co-founded Hive Spirit Technologies, Senior Researcher at NUS
Prague, Czech Republic
Hi, I’m David – co-founder of BottleCap AI, where we make large-language models 100 × faster and cheaper.

During my PhD I co-founded an AI-trading startup with Tomas Mikolov, applying language models to beat the markets, and published work at top conferences. Before that I pushed the boundaries of conversational AI as a finalist in Amazon’s Alexa Prize Grand Challenge.

News

3/3/2025 Launched BottleCap AI, aiming to revolutionize LLM efficiency. Featured in Forbes and CzechCrunch - 10M USD investment.
1/2/2025 Completed an internship at the National University of Singapore under Prof. Michael Shieh.
1/5/2024 Launched an algorithmic trading startup Hive Spirit Technologies LLC with Tomas Mikolov.
1/4/2024 Our paper Collapse of self-trained langauge models got accepted to ICLR 2024 as notable top-5% (Oral). (HackerNews discussion, arxiv)
5/3/2024 Interview for various news media about my language modeling research. (Forbes, CTK, Novinky, AI institute)
12/9/2023 We have won 3rd place in prestigious Alexa Prize Socialbot Grand Challenge 5 with team Alquist AI.
25/7/2023 Our paper on Preserving Semantics in Textual Adversarial Attacks got accepted to ECAI 2023. Notable-top-5% (Oral).
28/2/2021 Awarded with a GoodAI grant for the Emergence of Novelty research.

Notable publications

Collapse of self-trained language models David Herel and Tomas Mikolov In International Conference on Learning Representations, 2024 (notable).
ICLR 2024 (notable)
Alquist 5.0: Dialogue Trees Meet Generative Models. A Novel Approach for Enhancing SocialBot Conversations Ondrej Kobza, Jan Cuhel, Tommaso Gargiani, David Herel, Petr Marek, Jan Pichl, Jakub Konrad, Petr Lorenc and Jan Sedivy In Amazon Science, 2023.
Amazon Science 2023
On Difficulties of Attention Factorization through Shared Memory Uladzislau Yorsh, Martin Holeňa, Ondřej Bojar, David Herel In International Conference on Learning Representations, 2023.
ICLR 2023
Preserving Semantics in Textual Adversarial Attacks David Herel and Hugo Cisneros and Tomas Mikolov In European Conference on Artificial Intelligence Proceedings, 2023.
ECAI 2023
Thinking Tokens for Language Modeling David Herel and Tomas Mikolov In Artificial Intelligence and Theorem Proving Proceedings, 2023.
AITP 2023
Emergence of Novelty in Evolutionary Algorithms David Herel and Dominika Zogatova and Matej Kripner and Tomas Mikolov In Artificial Life Conference Proceedings, 2022.
ALIFE 2022