Roman Grebennikov
where else to find them
what they talk about
all 7 in the filters
where they keep going back
- Berlin Buzzwords2 editions2023–2025
- Scala.IO Nantes2 editions2016–2017
all 4 in the filters
on a stage with
who talks about the same things
- Pathikrit Bhowmick9 talks
- Chris Birchall5 talks
- Raimon Ràfols2 talks
- Jarek Ratajski25 talks
- Christophe Calvès4 talks
- Bill Venners21 talks
| title | conference | date |
|---|---|---|
| How [not] to evaluate your RAG | Berlin Buzzwords 2025 | 2025-06-15 |
A production RAG for a fintech support-agent assistant is built on a self-hosted Mixi/nixie search backend, with documents converted to Markdown and a… | ||
| Learning to hybrid search | Berlin Buzzwords 2023 | 2023-06-18 |
Uses the Amazon ESCI dataset and Metarank to compare BM25, LambdaMART LTR, sentence-transformer miniLM embeddings and cross-encoders; BM25/LTR sit around ~76… | ||
| Building an open-source ML feature store with Apache Flink | Flink Forward Global 2021 | 2021-10-18 |
An open source Flink-based framework called Featury implements a protobuf change log written to S3 that is readable both for low-latency online inference and… | ||
| CASE CLASSES ATE MY RAM | Scalar 2018 | 2018-04-06 |
Case-class instances and Scala collections often use huge hidden memory: tooling like Java Object Layout (JOL) and the jamm agent shows per-pageview overhead… | ||
| Case classes ate my RAM | Scala.IO 2017 | 2017-11-02 |
Case-class-heavy Scala heaps turn out to store tiny payloads and huge object overhead: a four-char String + long case class lands around ~72 bytes while naïve… | ||
| Is scala code slow? | Scala.IO 2016 | 2016-10-27 |
Claims are concrete: a one-line Scala rewrite can be measurably slower (author shows a 3× example) and small edits yield 4–7× gains. Demonstrations include JMH… | ||
| nothing matches | ||