Radu Gheorghe
where else to find them
no profile links on file
what they talk about
all 8 in the filters
where they keep going back
- Berlin Buzzwords6 editions2015–2026
on a stage with
who talks about the same things
- Philipp Krenn103 talks
- Mikkel Larsen3 talks
- Ruturaj Kadikar2 talks
- Aswath Srinivasan7 talks
- Solly Ross7 talks
- Jorge Turrado8 talks
| title | conference | date |
|---|---|---|
| Circular Dependency Fixes when Bootstrapping a Golden Set | Berlin Buzzwords 2026 | 2026-06-07 |
Workflow for bootstrapping a golden set that actually works: start from a 62K MovieLens sample enriched via TMDB, index into Vespa + navigator + Python… | ||
| Which GPU for Local LLMs? | Berlin Buzzwords 2025 | 2025-06-15 |
Benchmarks run with llama.cpp compare consumer and datacenter GPUs, showing H100 tops generation at ~300 tokens/s while an RTX 4090 often beats H100 on prompt… | ||
| Heap sizing and GC tuning for Solr and friends | Berlin Buzzwords 2024 | 2024-06-09 |
Solr/Elasticsearch-style nodes need heap sized from measured peaks not rules of thumb: take highest peak and lowest peak and use twice the low as a baseline… | ||
| Using TensorFlow in a Solr Query Parser | Berlin Buzzwords 2023 | 2023-06-18 |
A Python TensorFlow model does query tagging while BERT tokenization runs outside the model because TensorFlow for Java lacks tokenizer support; the SavedModel… | ||
| Autoscaling Elasticsearch for Logs on Kubernetes | Berlin Buzzwords 2022 | 2022-06-12 |
Explains a concrete autoscaling design for Elasticsearch/OpenSearch on Kubernetes: rotate time-based indices by size (recommended 10 GB per shard), update… | ||
| Autoscaling Elasticsearch for Logs on Kubernetes | KubeCon + CloudNativeCon Europe 2022 | 2022-05-16 |
An operator autoscaling Elasticsearch/OpenSearch for logs changes index templates to alter primary shard count, changes shards in template and rotates indices… | ||
| Using Neuroscience to Become a Better Developer | WebCamp Zagreb 2019 | 2019-10-11 |
Prefrontal cortex functions like a slow single-threaded CPU holding about 3–5 items of working memory and burning lots of energy, so plan to protect it: order… | ||
| Side by Side with Elasticsearch & Solr part 2 | Berlin Buzzwords 2015 | 2015-05-31 |
Benchmarks use JMeter hammering three m3.large-like VMs with Spanish Wikipedia, a 10M-video dataset and Apache logs; product search yields ~500 qps on… | ||
| nothing matches | ||