Almog Gavra

Before Responsive, Gavra worked on ksqlDB at Confluent and search infrastructure at LinkedIn. His professional portfolio also identifies OpenData, a collection of databases built on SlateDB, among his founding projects. His public education record lists Cornell University from 2012 to 2015.[1][4][5]

Keep the person in the code-review loop

tuicr presents a continuous diff in the terminal with Vim-style navigation. A reviewer can comment on a line, range, file or the whole change. It remembers reviewed files and hunks across sessions, then exports either a real review to a supported code host or structured Markdown that an agent can act on.[2]

Trade local-disk assumptions for object storage

SlateDB is an embedded key-value storage engine built as a log-structured merge tree. It stores sorted tables in object storage, using batched writes and caches to manage the higher request costs and latency. Gavra describes himself as a maintainer and contributor; the storage engine is a shared project, not his sole invention.[1][3]

Understand the machinery behind a state store

In a Kafka Summit presentation with Sophie Blee-Goldman, Gavra examines replacing built-in Kafka Streams state stores with an external database. The problems include exactly-once processing, preventing outdated writers from committing changes and preserving deletion markers. It connects a convenient storage interface with the distributed-system behavior that interface must uphold.[4]